No. 23-4361

No. 23-4362

Page 2Defendant – Appellant.

Appeals from the United States District Court for the Eastern District of Virginia, at Alexandria. Anthony John Trenga, Senior District Judge. (1:20-cr-00018-AJT-3; 1:20-cr00018-AJT-2; 1:20-cr-00018-AJT-5)

Argued: May 6, 2025 Decided: October 8, 2025 Amended: October 8, 2025

Before DIAZ, Chief Judge, NIEMEYER, Circuit Judge, and Matthew J. MADDOX, United States District Judge for the District of Maryland, sitting by designation.

Affirmed by published opinion. Judge Maddox wrote the opinion, in which Chief Judge Diaz and Judge Niemeyer joined.

ARGUED: Gregory Todd Hunter, Arlington, Virginia, for Appellants. Jacqueline Romy Bechara, OFFICE OF THE UNITED STATES ATTORNEY, Alexandria, Virginia, for Appellee. ON BRIEF: Adam M. Krischer, DENNIS, STEWART & KRISCHER, PLLC, Arlington, Virginia, for Appellant Jose Rosales Juarez. Thomas B. Walsh, PETROVICH & WALSH, PLC, Fairfax, Virginia, for Appellant Roberto Carlos Cruz Moreno. Jessica D. Aber, United States Attorney, Amanda St. Cyr, Assistant United States Attorney, OFFICE OF THE UNITED STATES ATTORNEY, Alexandria, Virginia, for Appellee.

Page 3MATTHEW MADDOX, District Judge:

Following a three-week jury trial, Kevin Perez Sandoval, Roberto Carlos Cruz Moreno, and Jose Rosales Juarez ('Appellants') were convicted of conspiracy to participate in a racketeering enterprise, conspiracy to commit murder in aid of racketeering, and conspiracy to distribute cocaine and marijuana, among other crimes. Perez Sandoval and Cruz Moreno were also convicted of attempted murder in aid of racketeering activity, assault with a dangerous weapon in aid of racketeering, and using and discharging a firearm during a crime of violence, and Rosales Juarez was convicted of being an accessory after the fact. In this appeal, Appellants argue that the district court abused its discretion in denying their motions for mistrial and for a new trial. Finding no abuse of discretion in the district court’s decisions, this Court AFFIRMS.¹

The evidence at trial established that Appellants were members of the Guancaos Lil Cycos Salvatruchas (“GLCS”), an MS-13 clique operating in northern Virginia. In March and August 2019, GLCS members attempted to murder two individuals. First, GLCS members lured E.P.A. into a car driven by Cruz Moreno, drove him into the woods, shot and stabbed him multiple times, attempted to slit his throat, and fled. GLCS members later planned the murder of N.M.S., with Rosales Juarez conducting surveillance of the victim.

Page 4Perez Sandoval drove GLCS members to retrieve a gun and changes of clothes, and dropped one GLCS member off so that he could sneak up to the victim and shoot him. After the conspirators fled, Rosales Juarez helped them evade law enforcement.

Appellants first challenge the district court’s denial of their motion for a mistrial based on issues raised about a court-appointed interpreter's translation during witness testimony. The district court's findings of fact are reviewed for clear error, and its conclusions of law are reviewed de novo. United States v. Roane , 378 F.3d 382, 395 (4th Cir. 2004). Based on this assessment of the underlying factual and legal issues, we decide whether the district court abused its discretion in denying the motion for mistrial. United States v. Wallace , 515 F.3d 327, 330 (4th Cir. 2008). “The denial of a defendant’s motion for mistrial is within the sound discretion of the district court and will be disturbed in only the ‘most extraordinary of circumstances.’” United States v. Beeman , 135 F.4th 139, 150 (4th Cir. 2025) (quoting United States v. Dorlouis , 107 F.3d 248, 257 (4th Cir. 1997)).

The district court appointed federally certified interpreters to provide language interpretation services for the trial, including Jose Lopez, Manuel Prado, Gregorio Ayala, and Victoria Dopazo. Two Spanish language interpreters were assigned for Appellants. Two other interpreters translated Spanish testimony into English for the jury, replacing or “spelling” each other at short intervals of 15 minutes or so.

Javier Bonilla, a former GLCS member involved in both attempted murders discussed during the trial, testified over three days. On the first day of Bonilla’s testimony, Page 5Lopez and Ayala translated for Appellants, and Prado and another interpreter translated for the witnesses, including Bonilla. The next day, Lopez and Ayala again translated for Appellants, and Prado and Dopazo translated for Bonilla. During cross-examination, Cruz Moreno's counsel questioned Bonilla about an April 2019 traffic stop, attempting to discern the location of a backpack containing a gun. Bonilla stated that the gun was in his backpack, that he was seated in the backseat of the car, and the backpack was on the floor.When questioned specifically about where he was sitting, he explained:

Q: Now, where were you seated in the car?

A: In the back seat, in the middle.

Q: That’s kind of – where your backpack was, right?

A: It wasn’t there. Q: Where was your backpack?

A: Somebody in the back had it. When the police came, they moved us.

J.A. 2372:22–2373:3.

At some point, the district court received a note from Juror 41 that read:

I am completely fluent in Spanish and the translation was really poor today. Some of the questions were not translated correctly. Another juror felt the same. [Prado] was the worst. It was tough at times. Don't think it always reflected the testimony.

Id. at 2503.

Appellants moved for a mistrial, and the district court conducted a hearing on the issue. The judge began by asking Juror 41 whether her concerns about the quality of the translation were confined to a single day, and she responded, “Yes, the bulk of it was today.

Page 6I mean, other days there were pieces where I was like, eh, that wasn’t exactly right, but somehow the answer came back okay.” Id. at 2478:23–2479:4. She later elaborated that the prior day’s interpretation was also confusing at times. When questioned about the note, she stated that the testimony that day was confusing because the witness would answer the question that the interpreter gave, which was not always the same as what the attorney asked, and at times the witness would jump in and correct the interpreter. When pressed for specific examples, Juror 41 recalled that she found the translation of Bonilla's testimony about what happened with the backpack containing the gun during the April 2019 traffic stop to be confusing. The other example she recalled was confusion over names when Bonilla was testifying about 'who was by the fence' when N.M.S. was shot.² Id. at 2484:14–2485:3.

Next, Lopez offered his opinion on Prado's interpretation, stating that the interpretation of Bonilla's testimony 'has been slightly shy of catastrophic.' Id. at 2486:21-2487:3. Lopez stated that it was difficult to understand Prado because he mumbled. When the court requested specifics, Lopez stated that the other interpreter corrected Prado several times, and he noted one instance of Prado flipping the meaning of a statement or phrase from positive to negative, but he could not recall the context. Lopez explained that, during their training, interpreters are taught to make corrections only when there has been a flipping of the meaning, “such as ‘I was not there’ versus ‘I was there,’”Page 7 and not to correct small stylistic choices. Id. at 2488:3-8. The court later pushed for clarification on what Lopez meant when he said that the interpretation was “slightly shy of catastrophic,” and Lopez explained that the interpretation “was both difficult to understand the pronunciation in English and in Spanish . . . and also inaccur[ate] in the translation itself, in the interpretation itself . . . .” Id. at 2496:1-15. He further stated that there were material interpretation issues with the meaning of how some things were translated but could not recall any examples.

The court then questioned Dopazo and Ayala. Dopazo was the backup interpreter for Prado during the second day of Bonilla’s testimony. She did not perceive any material misinterpretations, and she stated that she would have brought any material errors to the court’s attention. Ayala also did not perceive any material issues with Prado’s interpretation.

The district court also questioned Juror 66. Juror 66 explained that he was not fluent in Spanish but that he noticed certain words were not consistent and that Bonilla had corrected the interpreter. When pressed for specifics, Juror 66 remembered a misinterpretation of Bonilla’s username, javier_bonill13, where Prado interpreted 13 as three and Bonilla corrected him.

Before the judge ruled on the motion for mistrial, Lopez spoke up, stating that he recalled a 'severe misinterpretation.' Id. at 2534:16-20. Lopez recounted some of Bonilla’s testimony in Spanish and explained that a proper interpretation would be “[a] machete could be used to split a coconut open,” but the interpretation rendered was “[a] machete was used to open a box.” Id. at 2534:16-2535:1.

Page 8The district court denied the motion for mistrial but granted defense counsel leave to reexamine Bonilla about the issues identified by Juror 41. After recounting the evidence presented on the issue, the court stated that there was “no identification of any widespread issues” and found that the only specific interpretation problems related to discrete, limited issues involving one interpreter, who alternated with another interpreter, during one witness's testimony on one day. Id. at 2541:11-21. The court found that the limited instances before it did not “raise issues in a constitutional sense that would infect the entire fairness of the trial or really raise grave doubt as to the accuracy of all the other interpretations . . . .' Id. at 2542:17-25. No defense counsel took the opportunity to reexamine Bonilla after the district court’s ruling.

The federal courts have a program through which language interpreters are certified to provide services in court proceedings. See 28 U.S.C. § 1827. But even certified interpreters, being human, are not always perfect and may make mistakes. The question presented in this appeal is what level of language interpretation error calls for a mistrial.

Although this Court has not written extensively on this issue, the Eleventh Circuit recognized, citing other courts, that “the general standard for the adequate translation of trial proceedings requires continuous word for word translation of everything relating to the trial a defendant conversant in English would be privy to hear.” United States v. Joshi , 896 F.2d 1303, 1309 (11th Cir. 1990) (citing United States v. Lim , 794 F.2d 469, 470 (9th Cir.), cert. denied , 479 U.S. 937 (1986); United States v. Tapia , 631 F.2d 1207, 1209 (5th Cir. 1980); and United States v. Diaz Berrios , 441 F.2d 1125, 1127 (2d Cir. 1971)). ButPage 9 'minor deviations from this standard will not necessarily contravene a defendant's constitutional rights.' Id. 'The 'basic constitutional inquiry' when determining the competency of interpretation is ‘whether any in-adequacy in the interpretation made the trial fundamentally unfair.’” United States v. Garcia , 948 F.3d 789, 802 (7th Cir. 2020) (citation omitted); see also United States v. Mata , 181 F.3d 93 (4th Cir. 1999) (per curiam) ('The ultimate question is whether any inadequacy in the interpretation made the trial fundamentally unfair.” (quoting Valladares v. United States, 871 F.2d 1564, 1566 (11th Cir. 1989))). 'An interpretation need not be verbatim to be constitutionally sound if it reasonably conveys the intent or the idea of the thought spoken.” Garcia , 948 F.3d at 802 (citation and quotation omitted). 'Mere interpretation 'hiccups' do not create grave doubt[]” about the fundamental fairness of a trial. Id. ; see also Joshi , 896 F.2d at 1309 (“Defendant errs . . . in assuming that occasional lapses from [word for word translation], particularly when they are not objected to by the defendant, will render a trial fundamentally unfair.”). Importantly, “the trial court must be given ‘wide discretion’ in evaluating the adequacy of the interpreter's efforts.' Mata , 181 F.3d at 93 (quoting Valladares , 871 F.2d at 1566).

Appellants argue that there is clear and credible evidence that Prado’s interpretation was deficient, providing grave doubt that the testimony of a key government witness was accurate. According to Appellants, the interpretation errors violated their due process rights and interfered with their rights to confront the Government’s witness, and that these issues amounted to a structural error. The Government argues that the record does not reflect any widespread problems with Prado’s interpretation services and that the misinterpretations Page 10do not create grave doubt as to the overall accuracy of the interpretation during trial. Reviewing the language interpretation issues presented here, we conclude that they do not rise to the level of either a constitutional violation or a structural error.

“Structural errors are errors that affect the ‘entire conduct of the [proceeding] from beginning to end.’” Greer v. United States , 593 U.S. 503, 513 (2021) (quoting Arizona v. Fulminante , 499 U.S. 279, 309 (1991)); accord United States v. Ramirez-Castillo , 748 F.3d 205, 215–16 (4th Cir. 2014). Such errors “infect the entire trial process” and “necessarily render a trial fundamentally unfair,' Neder v. United States , 527 U.S. 1, 8 (1999) (quoting Brecht v. Abrahamson , 507 U.S. 619, 630 (1993), and then Rose v. Clark , 478 U.S. 570, 577 (1986)), and, accordingly, they are “subject to automatic reversal,” id. The category of structural errors is “very limited,” Ramirez-Castillo , 748 F.3d at 216 (quoting Johnson v. United States , 520 U.S. 461, 468 (1997)), and 'highly exceptional[,] . . . includ[ing], for example, the ‘denial of counsel of choice, denial of self-representation, denial of a public trial, and failure to convey to a jury that guilt must be proved beyond a reasonable doubt[,]’” Greer , 593 U.S. at 513 (quoting United States v. Davila , 569 U.S. 597 (2013)). “By contrast, discrete defects in the criminal process . . . are not structural because they do not ‘necessarily render a criminal trial fundamentally unfair or an unreliable vehicle for determining guilt or innocence.’” Id. (quoting Neder , 527 U.S. at 9).

We perceive no structural error here. What misinterpretations occurred did not “infect the entire trial process,” Brecht , 507 U.S. at 630, and did not fall within 'the very limited class of cases” where a structural error has been found, Johnson , 520 U.S. at 468. This Court has “cautioned that ‘before a court adds a new error to the list of structural Page 11errors . . . , the court must be certain that the error’s presence would render every such trial unfair[.]” Ramirez-Castillo , 748 F.3d at 216 (quoting Sherman v. Smith , 89 F.3d 1134, 1138 (4th Cir. 1996) (en banc)). Discrete and isolated language interpretation errors like those presented in this case would not render every trial involving comparable mistakes unfair.

Furthermore, the language interpretation issues did not render the trial in this case “fundamentally unfair.” Garcia , 948 F.3d at 802. The problems with Prado’s interpretation were limited to the testimony of one witness on one trial day, during which he switched with another interpreter about every 15 minutes. The language interpretation issues included unspecified errors in how Prado translated an attorney's questions and one unspecified instance in which, according to Lopez, Prado flipped the meaning of a statement or phrase. There is no evidence to suggest that any mistranslations of attorneys’ questions had any significant impact on the quality of the witness’s answers as conveyed to the jury. The fact that, at unspecified points during his testimony, the witness corrected the interpreter suggests that the translation of his testimony, by a certified Spanish interpreter, was accurate overall. There were also unspecified points at which another interpreter corrected Prado's interpretation. This fact, and Dopazo's statement that she would have brought any material interpretation errors to the district court’s attention, left the district court—and leaves this Court—with no reason to doubt that any interpretation errors by Prado that went uncorrected were minor.

The interpretation errors that were brought to the attention of the district court were indeed minor. These instances included an interpretation of Bonilla’s testimony about the Page 12backpack containing the gun, which Juror 41 found to be confusing; a mistranslation of a statement Bonilla made about a machete; a mistranslation in the number contained within Bonilla’s javier_bonill13 username; and an instance in which there was confusion over the name of a person who was by the fence at the time of the attempted murder of N.M.S.

Any confusion about the backpack containing the gun attributable to a translation error is immaterial. Bonilla’s testimony consistently confirmed that the backpack was in the backseat area of the car. Appellants offer no plausible argument that the location of the backpack was material to the jury’s guilty verdict on any count.

Similarly, any mistranslation of the number contained in the javier_bonill13 username was insignificant to the outcome of the trial. There was no factual dispute over the username, and Bonilla corrected the error.

Any error in the interpretation of Bonilla’s testimony regarding the name of a person by the fence during the shooting of N.M.S. was similarly immaterial. Bonilla consistently testified that a GLCS member who did not stand for trial in this case was alone when he shot N.M.S., after walking through the woods. The identity of the person by the fence when N.M.S. was shot had no bearing on any Appellant's responsibility for his role in the attempted murder or participation in the conspiracy.

Finally, Prado’s alleged mistranslation of Bonilla’s comment about a machete was not material. According to Lopez, Prado misinterpreted Bonilla’s statement, “A machete could be used to split a coconut open,” as, “A machete was used to open a box.” J.A. 2534:16–2535:1. Prado’s full translation was, “You could maybe use the machete for any number of things, maybe to open boxes.' Id. at 2396:10-11. Based on its context, Bonilla's Page 13comment appears to have been offered to explain that he could have possessed a machete for innocent purposes. Prado’s interpretation, while inaccurate, captured the gist and intent of the comment. It “reasonably convey[ed] the intent or the idea of the thought spoken,” Garcia , 948 F.3d at 802 (citation modified), and could not have prejudiced Appellants’ fair trial and confrontation rights.

We note Lopez's opinion that Prado's interpretation of Bonilla's testimony was “slightly shy of catastrophic.” J.A. 2486:21–25. But, when questioned about this opinion, Lopez focused on difficulty he had understanding Prado’s pronunciation. The only specific examples of inaccurate interpretation that Lopez could identify were the mistranslation of Bonilla’s statement about the machete and another instance in which Prado flipped the meaning of a statement, although Lopez could not recall the statement or its context. But two other interpreters—Dopazo and Ayala—stated that they did not witness any material errors in the accuracy of Prado’s interpretation. Importantly, Dopazo served as Prado’s backup interpreter on the trial day in question and stated that she would have notified the district court of any material errors in Prado’s interpretation if any occurred. Insofar as the district court credited the opinions of Dopazo and Ayala over Lopez’s opinion, we see no error or abuse of discretion in this determination.

In sum, the district judge did not err or abuse his discretion in concluding that the only specific language interpretation issues presented to him were “discrete” and “limited” in nature, did not create doubt as to the accuracy of Prado’s other interpretations, and did not compromise the fairness of the trial. Moreover, with respect to the potentially confusing testimony about the backpack with the gun, or the identity of the person by the fence, the Page 14district court adequately addressed any potential violation of Appellants' confrontation rights by giving defense counsel an opportunity to reexamine Bonilla about those matters.³ Defense counsel did not take that opportunity. Defense counsel’s decision not to reexamine Bonilla about those issues only lends further support to the conclusion that the issues were minor and any lapse in the accuracy of Prado’s translation of those points in Bonilla’s testimony does not create grave doubt as to the fundamental fairness of the trial. See Garcia , 948 F.3d at 802; Joshi , 896 F.2d at 1309.

We agree with the district court that the few identified instances of misinterpretation by Prado do not rise to the level of a constitutional violation. Because there was no constitutional violation, the district court did not abuse its discretion in denying Appellants’ motion for mistrial.

Appellants next challenge the district court’s denial of their motion for a new trial based on post-trial disclosure of unrelated official misconduct by the Government’s expert witness on MS-13. This Court reviews a district court’s denial of a motion for a new trial for abuse of discretion. United States v. Wolf , 860 F.3d 175, 189 (4th Cir. 2017). The district court's underlying legal determinations are reviewed de novo, and its factual findings are reviewed for clear error. Id. A district court abuses its discretion when it Page 15commits a legal error, and it would be legal error if the district court improperly determined that there was, or was not, a Brady violation. Id. (citing United States v. Bartko , 728 F.3d 327, 338 (4th Cir. 2013)).

On the second day of trial, the Government called Sergeant Claudio Saa of the Herndon Police Department (“HPD”) as an expert witness on Latino street gangs and MS13, specifically. During the trial, Sergeant Saa testified about MS-13 generally, including the different levels of gang membership, what the gang expected of its members, how members progress through the gang hierarchy, the gang’s violent tactics (including its use of machetes), and gang code. He testified that he had limited and outdated information regarding the GLCS clique of which Appellants were members. Sergeant Saa did not testify about any Appellant specifically or their conduct in connection with the charged offenses.

After the trial ended, the Government notified Appellants about allegations of official misconduct by Sergeant Saa. The Government disclosed that, at the time he testified, Sergeant Saa had been referred by HPD to the Federal Bureau of Investigation (the 'FBI') for gifting $29,000 worth of gun holsters to El Salvador's National Police (“PNC”) at the Salvadoran Embassy in December 2021, in his HPD uniform and without authorization. Before trial, members of HPD and an FBI public corruption squad met to discuss Sergeant Saa having potentially acted as an unregistered foreign agent in violation of federal law. The FBI then opened an investigation into Sergeant Saa’s conduct. The U.S. Attorney’s Office prosecuting the Appellants was made aware of the investigation after trial in this case concluded. The FBI ultimately determined that there was no evidence Page 16suggesting a crime had been committed or warranting further formal inquiry. After the FBI closed its investigation of Sergeant Saa’s conduct, HPD proceeded with its own internal investigation. HPD found that Sergeant Saa had violated several HPD policies. Sergeant Saa resigned while the investigation was still pending.

On August 29, 2022, the Government produced to Appellants, in redacted form, the FBI reports reflecting its investigation of Sergeant Saa. The next day, HPD produced a final memorandum summarizing its internal investigation. Separately, HPD requested that the Virginia Department of Criminal Justice Services (“VDCJS”) decertify Sergeant Saa as a law enforcement officer. VDCJS sustained that request.

The Government made its final disclosure to Appellants on November 25, 2022. This final disclosure included an appeal file from the VDCJS and an addendum prepared jointly by the FBI and the U.S. Attorney’s Office that responded to information in some of the HPD memoranda that “appeared to be inaccurate.” The Government also provided for the district court's in camera review unredacted copies of documents produced to Appellants.

Following the Government’s post-trial disclosure, Appellants filed a motion for a new trial, which the Government opposed, and the district court denied the motion in a written order. In its order, the district court assumed, without finding, that the Government suppressed the new evidence about Sergeant Saa, emphasizing that this assumption should not be construed as a finding of wrongdoing.

Yet, the district court determined that Appellants were not entitled to a new trial under Rule 33 of the Federal Rules of Criminal Procedure because the evidence was highly Page 17unlikely to produce an acquittal. The court found that Sergeant Saa’s testimony did not relate to Appellants specifically but instead concerned the structure and practices of MS13 generally. The court also determined that Sergeant Saa was, 'by no means, the Government's star witness,' noting that cooperating co-conspirators and victims also testified about MS-13 and the specific conduct at issue in the case. The court noted ample other evidence of Appellants’ guilt as well, including physical evidence, phone records, and Appellants’ own statements. The court found all of this evidence sufficient to support the convictions without Sergeant Saa's testimony. The court also rejected Appellants' argument that the new evidence about Sergeant Saa uncovered a conspiracy between MS13, Commissioner Funes, and Sergeant Saa to convict Appellants. The court found this theory to be implausible and unsupported, and that it would not likely produce an acquittal.

The district court also determined that Appellants were not entitled to a new trial based on a Brady violation. The court found “no reasonable probability” that a disclosure of the information about Sergeant Saa before trial would have produced a different result. Even assuming that the evidence was improperly suppressed, the court found, the nature of the evidence did not undermine confidence in the outcome of the trial because there was sufficient evidence to convict, even excluding Sergeant Saa’s testimony. Relatedly, the district court rejected Appellants’ Confrontation Clause argument because any error in their lost opportunity to cross-examine Sergeant Saa about the matters disclosed after trial was harmless beyond a reasonable doubt.

Rule 33 of the Federal Rules of Criminal Procedure authorizes a district court to “vacate any judgment and grant a new trial if the interest of justice so requires.” Fed. R. Crim. P. 33(a). When reviewing a motion for new trial under Rule 33, courts analyze five factors:

(a) the evidence must be, in fact, newly discovered, i.e., discovered since the trial; (b) facts must be alleged from which the court may infer diligence on the part of the movant; (c) the evidence relied on must not be merely cumulative or impeaching; (d) it must be material to the issues involved; and (e) it must be such, and of such nature, as that, on a new trial, the newly discovered evidence would probably produce an acquittal.

United States v. Banks , 104 F.4th 496, 509 (4th Cir.), cert. denied sub nom. Davis v. United States , 145 S. Ct. 338 (2024), and cert. denied sub nom. Anderson v. United States , 145 S. Ct. 563 (2024) (quoting United States v. Robinson , 627 F.3d 941, 948 (4th Cir. 2010)).

Appellants’ bid for a new trial fails under the Rule 33 standard. The information disclosed about Sergeant Saa after trial was that he that had gifted gun holsters to a Salvadoran law enforcement agency, he was investigated for having potentially acted as an unregistered foreign agent, and he was found to have violated HPD policies, which later resulted in his decertification as a law enforcement officer. None of this information had any bearing on the facts of this case. At best, it was grounds for impeaching Sergeant Saa’s testimony.

As a general matter, newly discovered evidence that is “merely impeaching” does not warrant a new trial. Banks , 104 F.4th at 511 (quoting United States v. Custis , 988 F.2dPage 19 1355, 1359 (4th Cir. 1993)). “[M]otions for a new trial based on impeaching evidence discovered after trial should be granted 'only with great caution and in the most extraordinary circumstances,’” where there is “‘a real concern that an innocent person may have been convicted.’” Custis , 988 F.2d at 1360 (quoting United States v. Sanchez , 969 F.2d 1409, 1414 (2d Cir. 1992)), quoted in Banks , 104 F.4th at 511. Newly discovered impeachment evidence may suffice to merit a new trial under Rule 33 if it impeaches “the uncorroborated testimony of a single witness who was discovered after trial to be utterly unworthy of being believed[,]' and 'the government's case rested entirely on [that] testimony . . . .” Custis , 988 F.2d at 1359 (quoting United States v. Taglia , 922 F.2d 413, 415 (7th Cir. 1991)).

Here, Sergeant Saa’s trial testimony was not uncorroborated. Sergeant Saa was one of over 30 witnesses for the prosecution. His testimony concerned the structure and practices of MS-13 in general, not the specific conduct in this case, and was corroborated by the testimony of MS-13 members and Appellants’ own statements to law enforcement. The Government's case rested neither entirely nor substantially on Sergeant Saa's testimony. Therefore, the district court did not abuse its discretion in denying Appellants’ motion for new trial under Rule 33.

Under Brady and Giglio , “a defendant’s due-process rights are violated when the prosecution suppresses evidence ‘favorable to an accused’ or for impeachment purposes.” Banks , 104 F.4th at 509 (citing Brady v. Maryland , 373 U.S. 83, 87 (1963) and Giglio v. United States , 405 U.S. 150 (1972)). When a defendant asserts a Brady / Giglio violation, a Page 20defendant may obtain a new trial if he can show that “(1) the undisclosed evidence was favorable to him; (2) it was material, i.e., caused him prejudice; and (3) the prosecution possessed the evidence, yet failed to disclose it.” Id. (citing United States v. Stokes , 261 F.3d 496, 502 (4th Cir. 2001)). “[F]avorable evidence is material, and constitutional error results from its suppression by the government, ‘if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.’” Kyles v. Whitley , 514 U.S. 419, 433–34 (1995) (quoting United States v. Bagley , 473 U.S. 667, 682 (1985)). The Supreme Court explained in Kyles :

[A] showing of materiality does not require demonstration by a preponderance that disclosure of the suppressed evidence would have resulted ultimately in the defendant’s acquittal . . . . Bagley ’s touchstone of materiality is a “reasonable probability” of a different result . . . . The question is not whether the defendant would more likely than not have received a different verdict with the evidence, but whether in its absence he received a fair trial, understood as a trial resulting in a verdict worthy of confidence. A 'reasonable probability” of a different result is accordingly shown when the government’s evidentiary suppression “undermines confidence in the outcome of the trial.”

Id. at 434 (quoting Bagley , 473 U.S. at 678). Stated differently, “the question is whether the favorable evidence, considered collectively, could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict .' United States v. George , 95 F.4th 200, 209 (4th Cir. 2024) (citing United States v. Blankenship , 19 F.4th 685, 692 (4th Cir. 2021)).

This Court has held that “the materiality requirement is not satisfied when a law enforcement officer’s misconduct is tangential to the evidence establishing a defendant’s Page 21own culpability.” Banks , 104 F.4th at 511. In Robinson , for example, evidence of a law enforcement witness’s official misconduct unrelated to the case at bar was insufficient to warrant a new trial where testimony from numerous other law enforcement witnesses, codefendants, and cooperating witnesses and ample physical evidence established the defendant's guilt. 627 F.3d at 952-53. See also Wolf , 860 F.3d at 192 (holding that defendant could not demonstrate prejudice in light of “overwhelming evidence of guilt”); United States v. Bartko , 728 F.3d 327, 340-41 (4th Cir. 2013) (upholding trial court determination that impeachment evidence was not material where there was a “mountain of evidence marshaled against [the defendant]”).

The post-trial disclosure at issue here concerns Sergeant Saa having been investigated and found to have engaged in official misconduct by gifting gun holsters to a Salvadoran law enforcement agency. Viewing this information against the evidence of each Appellant’s guilt presented at trial, we cannot conclude that it “put[s] the whole case in such a different light as to undermine confidence in the verdict.” George , 95 F.4th at 209 (citation omitted).

While Sergeant Saa was the only law enforcement official who testified generally about MS-13’s structure and tactics, he was not the prosecution’s star witness on this topic, nor any other topic. Four MS-13 cooperators testified about the structure and organization of the gang, and that testimony was corroborated by Appellants’ own statements. Three cooperators, also members of GLCS, testified about the gang's rank structure, rules, common terminology, code words, and operations. Further, wiretap conversations, messages and pictures exchanged between co-conspirators, video recordings, and law Page 22enforcement interviews with Appellants independently supported the cooperators’ testimonies.

We find no error in the district court’s conclusions that evidence of Appellants’ guilt from sources other than Sergeant Saa was overwhelming, there was 'no reasonable probability of a different result,” and the nature of the official misconduct evidence “does not undermine the confidence in the outcome of the trial.” J.A. 3304–05. Nor is there a reasonable probability that the information about Sergeant Saa's unrelated official misconduct disclosed after trial would have produced a different result if it had been disclosed before trial. See Kyles , 514 U.S. at 433–34.

This Court has held that evidence is material when 'the question of guilt or innocence essentially involved a credibility determination between the defendant and a witness” and the evidence “seriously undermine[s]” the witness’s credibility. George , 95 F.4th at 209 (quoting United States v. Kelly , 35 F.3d 929, 937 (4th Cir. 1994)). “At the same time, however, where the withheld evidence is impeaching but the witness was impeached in other ways at trial, we have held that the withheld evidence was not material.” Id. (citing United States v. Hoyte , 51 F.3d 1239, 1243 (4th Cir. 1995), and United States v. Chavez , 894 F.3d 593, 600 (4th Cir. 2018)). Here, Sergeant Saa was impeached during voir dire and cross-examination by defense counsel. He conceded that he had not personally participated in a gang investigation for nearly ten years before testifying, he had no advanced degrees relating to gangs, he did not study current literature on the topic, he had not taken any recent academic courses on MS-13, his information regarding MS-13’s rules was dated, he only had general knowledge of GLCS, and he had not investigated Page 23Appellants. The post-trial disclosure did not deprive Appellants of their opportunity impeach Sergeant Saa as to his qualifications and the foundations for his testimony as an expert. The official misconduct evidence disclosed after the trial in this case was not material.

Appellants argue that the post-trial disclosure prejudiced their rights under the Confrontation Clause. This claim is subject to review for harmless error. See United States v. Abu Ali , 528 F.3d 210 (4th Cir. 2008) (citing Delaware v. Van Arsdall , 475 U.S. 673, 684 (1986)). “The correct inquiry is whether, assuming that the damaging potential of the cross-examination were fully realized, a reviewing court might nonetheless say that the error was harmless beyond a reasonable doubt.” Van Arsdall , 475 U.S. at 684. Considering the voluminous evidence of Appellants’ guilt and the opportunities they had to impeach Sergeant Saa as to his qualifications, any erroneous deprivation of Appellants’ opportunity to cross-examine Sergeant Saa specifically about the allegations and investigation of his unrelated official misconduct was harmless beyond a reasonable doubt.

Because there was overwhelming trial evidence presented to establish Appellants’ guilt, there was no reasonable probability of a different result had the post-trial disclosure about Sergeant Saa been made before trial, and Appellants were capable of impeaching Sergeant Saa without the information disclosed after trial, we conclude that the district court did not err in finding that the new evidence was not material under Brady and Giglio and denying the motion for new trial.

For the reasons set forth above, the judgment of the district court is Page 24AFFIRMED .