Tsolak Gevorgyan was convicted of healthcare fraud and violating the Anti-Kickback Statute, including conspiracy and substantive counts. We AFFIRM these convictions.
I.
In 2006, Leonard Kibert, a physician, established the New Life Sleeping and Allergy Disorder Center ("the Clinic") in Houston, Texas. Gevorgyan was employed by the Clinic as an "office manager" and "operations person," where he cut a ubiquitous figure. There were no formal job titles at the Clinic, but Gevorgyan was described as "kind of r[unning] the [C]linic" on a day-to-day basis. In particular, Gevorgyan paid rent, faxed documents, *453met with patients, conducted sleep and allergy tests, and occasionally participated in job interviews with prospective employees.
That the Clinic's practice was sustained in part by illegal activity is not at issue in this appeal. Of relevance here, two distinct sets of practices befouled the Clinic. First, the Clinic engaged in the payment of "marketers" who themselves paid patients to attend the Clinic. Second, the Clinic routinely sought and received reimbursement for medical tests it never conducted.
Gevorgyan was involved in each of these schemes. With respect to the payment of kickbacks for patients, Gevorgyan personally paid a marketer named Leon, who then distributed the money to the other marketers who succeeded in bringing patients to the Clinic. At some point during the Clinic's lifecycle, one of Gevorgyan's erstwhile codefendants, Robert Manning, replaced Leon in his role as middleman. As before, Manning paid the marketers, who themselves provided money to prospective patients; at the end of the week, Gevorgyan reconciled a ledger with the number of patients who had visited the Clinic to the number that Manning had recorded based on his own interactions with marketers. Gevorgyan then paid Manning based on the number of patients brought in. The Clinic's patient base was unique; the Clinic only served Medicare beneficiaries, and it also apparently never charged copays.
Gevorgyan gave checks to Manning, payable to a company named GGM Inc., which Manning had created for the purpose of holding the money he received from the Clinic. Similarly, the checks provided by Gevorgyan did not draw upon the accounts of Gevorgyan or the Clinic; they were drawn from the account of a corporation called Gohev Services, which by corporate filings Gevorgyan served as an owner, officer, and director. These checks were signed by Arsen Gevorgyan, Gevorgyan's older brother. Gohev Services was also "intertwined" in the Clinic's general activities, financing many of the day-to-day costs incurred by the business.
With respect to the healthcare fraud, in addition to "kind of r[unning] the [C]linic" in the ways described above, Gevorgyan was also involved in the intake of new patients and had apparently trained the Clinic's receptionist on how to register them.
In August 2009, Kibert closed the Clinic. Several years later, the Texas government began investigating and interviewing actors associated with it on the basis of tips from marketers, which led to the indictment of Gevorgyan alongside Kibert, Manning, and two other codefendants. The indictment charged all of the defendants with conspiracy to commit healthcare fraud and various individual counts of healthcare fraud; it charged Gevorgyan and Manning with conspiracy to violate the Anti-Kickback Statute and five counts of unlawfully making kickback payments. Manning pleaded guilty to his charges and provided testimony against his codefendants. Following a 14-day trial, Kibert and Gevorgyan were convicted while the other two remaining codefendants were acquitted.1 On January 11, 2017, Gevorgyan was sentenced to 58 months of imprisonment followed by three years of supervised release. Gevorgyan appealed, challenging both the sufficiency of the evidence supporting his convictions and the district court's failure to sever him from the trial.
*454II.
In review of the sufficiency of evidence underlying criminal convictions, we ask whether "a rational jury could have found each essential element of the offense beyond a reasonable doubt."2 All evidence must be taken in the light "most favorable to the verdict and drawing all reasonable inferences from the evidence to support the verdict."3 Gevorgyan challenges both sets of his convictions on the basis of insufficient evidence: first, he challenges his convictions for conspiracy to violate, and substantive violation of, the Anti-Kickback Statute; second, he challenges his convictions for conspiracy to commit, and substantive commission of, healthcare fraud. We take these two categories of challenges in turn.
A.
Gevorgyan challenges the sufficiency of the evidence underlying his Anti-Kickback Statute convictions. The Anti-Kickback Statute in effect during the relevant time period proscribed "knowingly and willfully offer[ing] or pay[ing] any remuneration ... directly or indirectly, overtly or covertly, in cash or in kind to any person to induce such person to refer an individual to a person for the furnishing or arranging for the furnishing of any item or service for which payment may be made in whole or in part under a Federal health care program."4 A conspiracy to violate the Anti-Kickback Statute under 18 U.S.C. § 371 requires "an agreement to do so, knowing and voluntary participation in the conspiracy, and an overt act by one member in furtherance of the unlawful goal."5 Gevorgyan focuses only on the mens rea components of these offenses, arguing that he did not "knowing[ly] and willing[ly]" violate the statute or participate in any conspiracy.
The government claims that Gevorgyan's challenge to the Anti-Kickback Statute convictions was not properly preserved, asserting that, at the close of evidence, Gevorgyan only reurged his challenge to the evidence presented in support of healthcare fraud, not the Anti-Kickback Statute counts. If Gevorgyan abandoned his challenge to the evidence underlying the Anti-Kickback Statute counts, we would review the imposition of the convictions with greater deference to the district court; we would ask whether the convictions impose a "manifest miscarriage of justice," which requires Gevorgyan to demonstrate that "the record is devoid of evidence pointing to guilt" or that "the evidence on a key element of the offense is so tenuous that a conviction would be shocking."6
This we need not do, for even if we decided that the objections were fully preserved, we would uphold Gevorgyan's convictions. On this record, and drawing all reasonable inferences in support of the verdict, we must conclude that the jury could have found that Gevorgyan was aware of the nature and purpose of the kickback payments he made to Manning.
*455Gevorgyan understandably attacks the testimony of Manning, his former coconspirator-he argues that Manning was an inherently unreliable witness given his own role in the scheme, and points out that there are alternative explanations for the payments, including that they were simply for "crowd control" or "keep[ing] some semblance of order at the clinic." Perhaps this is so, but the fact remains that there is sufficient evidence to support the jury's conclusion that Gevorgyan knew the payments were kickbacks. Manning's testimony aside, evidence indicates that Gevorgyan instructed the Clinic's receptionist to keep a ledger tallying the number of incoming patients. This patient ledger contained total figures written by Gevorgyan that directly mirrored the figures on the payments that Gevorgyan delivered to Manning. And Gevorgyan served as one of the owners of the corporation that acted as the source of those payments.
This said, there is no reason for us to discount Manning's testimony. Manning directly testified about his conversations with Gevorgyan, explaining that Gevorgyan assigned his particular role in the scheme and opining that Gevorgyan knew that the purpose of the payments was for marketers to bring patients in. Manning openly testified that he had entered into a plea agreement and hoped to receive a reduced sentence in light of his cooperation with the government, and the district court appropriately instructed the jury that the testimony of Manning should be "received with caution and weighed with great care." Weighing Manning's credibility in determining Gevorgyan's knowledge of the kickbacks was the appropriate task of the jury.7
Gevorgyan's sufficiency of the evidence challenge to his conspiracy and substantive convictions for violation of the Anti-Kickback Statute must fail.
B.
Gevorgyan also challenges the sufficiency of evidence underlying his healthcare fraud convictions. To be guilty of healthcare fraud under 18 U.S.C. § 1347, a defendant must have "knowingly and willfully executed a scheme to defraud a government health care program like Medicare."8 Further, under 18 U.S.C. § 1349, "[w]illfully joining with another to engage in such a scheme, with awareness of the agreement's unlawful purpose, is a conspiracy to commit healthcare fraud."9 As with his challenge to the Anti-Kickback Statute violations, Gevorgyan takes aim only at the mens rea requirements in these offenses, namely that he knew of the Clinic's ongoing healthcare fraud; he claims that the government failed to show that he "knowingly and willfully" committed healthcare fraud and, with respect to any conspiracy to do so, that he was not "aware[ ] of the agreement's purpose."
Gevorgyan points to his relative youth and inexperience in the medical field, arguing that these features rendered any knowledge of the Clinic's healthcare fraud unlikely. He also pushes back against the evidence produced at trial characterizing him as a manager of the Clinic-instead, he argues that he was only *456a low-level worker and unwitting participant in the Clinic's fraud.
His arguments here have more purchase than with respect to the Anti-Kickback Statute violations, but they nonetheless fail. There was sufficient evidence to support the jury's determination that Gevorgyan knew of the Clinic's healthcare fraud. First, the jury heard testimony that Gevorgyan trained and supervised others in performing patient intake, which suggests that he directly knew of the Clinic's practice of only accepting Medicare patients and of not accepting copayments. These are "unusual practice[s]" that are probative of intent to defraud.10 Second, Gevorgyan's central involvement in the kickback scheme alongside Manning supported the related inference that he knew about the healthcare fraud; the marketers receiving kickbacks, after all, recruited patients with no regard for their actual need for medical services.11 Third, Gevorgyan was an owner, director, and officer of Gohev Services. Evidence indicates that Gohev Services served as the repository of the fruits of the Clinic's healthcare fraud, including eventually distributing approximately $119,000 of the funds to Gevorgyan over several years.12 He claims that because he never signed the underlying corporate filings or personally opened any bank account for the firm, the jury could have decided that he was unaware of these roles. But taken in the light most favorable to the verdict, we cannot conclude that it erred in inferring that he was aware of his position and the profit to which it entitled him. Finally, the jury heard ample evidence that Gevorgyan did serve in a management role at the Clinic, despite Gevorgyan's continued protestations otherwise.13
All of the foregoing facts taken together support the jury's determination that Gevorgyan knew of the Clinic's healthcare fraud. We decline to vacate Gevorgyan's convictions for commission of, or conspiracy to commit, healthcare fraud.
III.
Gevorgyan argues that the district court erred in failing to sever him from his codefendants at trial. Ordinarily, a district court's failure to sever a defendant from a trial is reviewed for abuse of discretion.14 However, when a defendant fails to timely assert the grounds upon which he seeks severance, plain error applies.15
In this case, Gevorgyan's severance argument rests on two different theories: first, he claims that a disparity of evidence existed between him and his codefendants; second, he claims that he suffered anti-Armenian bias stoked by one codefendant *457during trial. Gevorgyan only moved for severance from his codefendants on the basis of the two theories he now pursues after his trial concluded, and this in the context of a motion for a new trial.16 For this reason, and as Gevorgyan conceded during oral argument, the plain error framework here guides our review.17 That framework requires that the district court's decision not to sever Gevorgyan from his codefendants be a result of "a clear or obvious error, which affected [Gevorgyan's] substantial rights, and which would seriously affect the fairness, integrity or public reputation of judicial proceedings if allowed to stand" before we can overturn it.18
No plain error exists in this case. There is little reason to believe that the district court's failure to sever Gevorgyan affected his "substantial rights," much less satisfied the other plain error requirements. As a substantive matter, "this [C]ourt has been reluctant to vacate a conviction because the district court refused to sever a trial."19 In order to justify vacating a conviction, therefore, "the defendant must show specific and compelling prejudice resulting from the joint trial," by both "isolat[ing] events occurring in the course of the trial" and "demonstrat[ing] that such events caused substantial prejudice."20
Gevorgyan's first theory-that a disparity of evidence existed between him and his codefendants, unfairly prejudicing him before the jury-rests on a faulty reading of the record. Ample evidence exists tying Gevorgyan to his charged offenses, as we have outlined above. Further, we have previously held that provision of appropriate jury instructions cuts against any spillover prejudice from failure to sever,21 and we have also held that a jury's decision to acquit other codefendants provides evidence of its ability to make evaluations on an appropriately individualized basis.22 Both factors are present here: the district court provided clear instructions dictating that the jury consider all defendants individually, and the jury indeed acquitted one of Gevorgyan's codefendants of all charges. In light of these facts, we cannot agree that a disparity of evidence affected Gevorgyan's substantial rights.
Gevorgyan's second theory of plain error must also fail. He recounts various references made during the trial to "Armenians"-mostly elicited by Kibert-that he claims, in the aggregate, biased the jury against him by stoking anti-Armenian prejudice. He points to two trial components *458as particularly problematic: first, an IRS agent testified about Kibert's prior statements to law enforcement, which Gevorgyan claims included "prejudicial statements" about Armenians; and second, Kibert purportedly made a defense strategy of blaming "an Armenian or Russian criminal enterprise" for the Clinic's legal woes.
Gevorgyan simply overreads both the record and the significance of the utterances he points to. The IRS agent's testimony recounted Kibert's earlier, generalized statements explaining his interactions with a group of Armenians, but these statements are flatly descriptive and are not reasonably read as ethnically inflammatory. Additionally, the district court instructed the jury not to consider them with respect to Gevorgyan for other evidentiary purposes, and there is no reason to believe that the jury did not obey this instruction.23
Gevorgyan attempts to cast other statements made by Kibert as parts of an overarching anti-Armenian defense. This characterization lacks support. In fact, the record reveals that the district court was attuned to the possibility of anti-Armenian prejudice, and it worked to avoid any chance that ethnic innuendo could taint the proceedings. The references to Armenians that did enter the record were, once more, essentially flat descriptions of actors purportedly involved in the Clinic. They certainly did not suggest any latent criminality on the part of Armenians in general. Indeed, all of the inculpatory evidence tying Gevorgyan to the Clinic's unlawful schemes had nothing to do with his identity as an Armenian and everything to do with his substantive involvement at the Clinic. The district court's decision not to sever Gevorgyan on the basis of his identity as an Armenian did not affect his substantial rights, and the district court therefore did not commit plain error.24
IV.
For the foregoing reasons, we AFFIRM Gevorgyan's convictions and the district court's rulings below.