AKASA Seeks Rare Hybrid Talent as Healthcare AI Automation Scales

Where AKASA Is Hiring Now

AKASA has 9–10 open roles across engineering, product, and compliance, with six in engineering, two in product, one each in compliance, security, and a chief of staff — signaling a platform-scale bet on generative AI for healthcare revenue cycle automation. The expansion reflects surging enterprise demand for AI-driven claims and prior-authorization tools, triggering intense competition for talent with rare dual expertise in hospital operations and large language model application.

Engineering dominates, with six of nine board roles as of mid-August (two added in the past week per Zero G Talent), and the split reveals where the platform is thickening. Two senior backend slots (one in each hub), a remote ML hire, a DevOps role in San Francisco, and two client-facing engineering tracks signal simultaneous investment in core model infrastructure, deployment reliability, and the "expert-in-the-loop" delivery model AKASA sells to health systems. The Forward Deployed Engineer role, posted in both New York and San Francisco, is the clearest proxy for implementation capacity: these engineers sit with customers to tune generative AI agents against live prior-auth and claims workflows.

Geographically, the Bay Area holds five roles across San Francisco and South San Francisco; New York carries three; one is remote with U.S. location restrictions. AKASA's careers page describes a "remote-friendly, hybrid organization with team members in 29 states" and hubs in the Bay Area, New York, and Denver — yet Denver shows no open roles. The concentration in South San Francisco (three roles, including the Director of Compliance & AI Governance and the Senior TPM) suggests the compliance and program-management layer is anchoring near headquarters, while New York's Chief of Staff and two senior engineering hires point to a second center of gravity for product strategy and East Coast client deployment.

The Product Operations & QA Specialist at $60k–$85k, according to JobScroller, stands out as the only sub-$100k role — a tactical hire for release validation and customer-facing quality checks. Its persistence (posted three weeks ago per JobScroller) may indicate a narrower candidate pool for healthcare-specific QA than for engineering. Meanwhile, JobScroller's figures put that role at $150k–$185k (a role that barely existed industry-wide two years ago), reflecting the regulatory premium on HIPAA-fluent AI oversight.

What the distribution doesn't show is equally telling: no dedicated clinical or revenue-cycle subject-matter-expert roles, no sales or customer-success listings, no Denver postings despite the named hub. AKASA appears to be hiring the builders and the governance layer first, trusting its existing clinical-operations bench or its contractor network to carry domain depth.

The Screen: Revenue Cycle Fluency First

AKASA's job descriptions read less like typical AI company postings and more like revenue cycle operations manuals with a machine learning overlay. The screening bar is explicit: candidates who cannot demonstrate hands-on fluency in the day-to-day mechanics of hospital revenue cycle (eligibility verification, prior authorization workflows, claim edit resolution, UB-04 and 1500 form preparation, denial appeals, secondary and tertiary insurance coordination) do not advance, regardless of their model-tuning pedigree.

The Revenue Cycle Operations Projects Lead role requires four to seven years as an analyst in an RCM or patient access environment, or in RCM consulting or project management. That experience must cover patient registration and financial clearance (insurance verification, cost estimates, prior authorizations) as well as claims billing, claim edits and submissions, and claims follow-up. The posting adds a hard floor: at least three years with Epic, Cerner, or another major EHR, and a preference for multi-system exposure. A Revenue Cycle Operations Analyst role echoes the same requirement: three-plus years of Epic, Cerner, or equivalent EHR experience, plus direct familiarity with obtaining pre-authorizations, checking eligibility and benefits, preparing and reviewing 1500 and UB-04 claims, transmitting claims electronically or on paper, auditing insurance payments against contractual discounts, following up on unpaid and underpaid claims through rebills and appeals, identifying and billing secondary and tertiary insurers, and reviewing accounts for patient follow-up.

These are not "nice to have" qualifications. They are the filter. AKASA's platform automates revenue cycle management using computer vision-based RPA, machine learning, and human-in-the-loop workflows that analyze clinical and administrative documents, extract data, apply ML decisions, and execute robotic automation inside healthcare IT systems — with humans reviewing uncertain outcomes to ensure accuracy. The company trains health-system-specific models rather than relying on generic foundations; its platform outperforms GPT-4 by as much as 40 percent on revenue-cycle tasks because it learns from each client's unique data. That approach only works if the people building, deploying, and monitoring the automation understand the workflows they are replacing at the keystroke level.

The hybrid requirement shows up in the day-to-day responsibilities. The RCO Projects Lead "provides revenue cycle subject matter expertise across the company and represents RCO in new product development, existing product implementations, and new client conversations and projects." They "work cross-functionally across Sales, Marketing, and Engineering to deepen the revenue cycle expertise within the company," conduct "process mapping and shadowing with clients to understand workflows," and "develop Revenue Cycle training specification documents." The analyst role explicitly states: "Utilize data labeling of payer portals, EHR data, and payer responses to train our AI model" and "Draw upon past experience in billing, coding, and follow-up to advance patient accounts on behalf of our hospital clients."

Engineering roles carry the same dual mandate. The board lists a Senior Machine Learning Engineer (remote) and a Senior Forward Deployed Engineer (New York City) among recent openings. Forward deployed engineers at AKASA sit at client sites (health systems), translating clinical and administrative workflows into model inputs and automation logic. That role cannot be filled by a pure ML researcher; it demands someone who can read an Epic build sheet, trace a prior auth denial to a missing CPT code, and then design the labeling schema that teaches the model to catch it next time.

The compensation bands reflect the scarcity. Zero G Talent's data shows senior backend and DevOps engineers in South San Francisco and San Francisco command the top of the board's range. JobScroller's figures put a Senior Technical Program Manager in South San Francisco near the median. JobScroller found an IT Security Operations Engineer in San Francisco sits at the lower third. These are not premium AI salaries; they are premium hybrid salaries. The market pays for the intersection, not the parts.

Cleveland Clinic's expansion of AKASA after a successful pilot and its 2025 deployment of AKASA coding and CDI tools across U.S. locations — validates the model. Health systems demand proof; one weak deployment damages enterprise credibility fast. FDA, HIPAA, and claims-audit scrutiny on LLM-generated coding can trigger contract exits. AKASA's screening criteria are a direct response to that pressure: every hire must be able to defend the automation's output in the language of the revenue cycle, not just the language of the model.

Why the Revenue Cycle Layer Is Wide Open

The numbers are staggering. Total U.S. healthcare administration spending hits $740 billion annually, yet healthcare IT captures just $63 billion — a penetration rate below 9%. Medical documentation and back-office revenue cycle management together represent nearly 60% of all healthcare IT spending, a combined market of roughly $38 billion.

Providers are driving this. They supply three-quarters of the $1.4 billion now flowing into healthcare AI. National health expenditures grew 7.2% to $5.3 trillion in 2024, reaching 18% of GDP, and CMS projects that share climbing to 20.6% by 2034. Administrative overhead continues to erode margins and burn out clinicians, compounding post-pandemic labor shortages. Claim denials are rising. Billing complexity is growing. Prior authorization remains healthcare's most reviled administrative process and the second most time-consuming RCM task.

AKASA sits squarely in the administrative workflow layer — not clinical coding, not payment posting, but the 40% of revenue cycle staff time spent on eligibility checks, prior auth status calls, claim status inquiries, and denial follow-up. That layer represents $98 billion in annual administrative services where software penetration is only 3%. Patient engagement and access account for more than $100 billion in admin spending with 5% software capture. The AI market addressing these workflows has already topped $100 million and is growing 10x year over year.

Client demand is concrete. Schneck Medical Center cut claim denials 4.6% monthly within six months of deploying Experian Health's AI Advantage platform; average claim processing time dropped from 12–15 minutes to under five. Advocate Health evaluated over 225 AI solutions and selected 40 use cases, including the largest Microsoft Dragon Copilot deployment plus imaging tools from Aidoc and Rad AI and AI for call centers — projected to cut documentation time by more than half while automating prior authorizations, referrals, and coding. Mayo Clinic is investing more than $1 billion across 200-plus AI projects. SimonMed scaled from co-building with fewer than 10 vendors to piloting 50-plus.

The competitive dynamics are sharpening. Startups capture 85% of generative AI spend in healthcare, but most customers say they prefer buying AI from their incumbent EHR. Epic, Oracle Health, and athenahealth have all launched ambient scribes and are embedding AI directly into their platforms. For coding, billing, prior authorization, scheduling, clinical decision support, and patient navigation, the survey data shows clear EHR preference. Meanwhile, payers are watching with growing concern. They fear surging call and claims volume from AI-enabled provider tools that let hospitals submit claims instantly, check status, and file appeals — automation that threatens to overwhelm payer call centers. They also worry about coding tools optimized to maximize reimbursement. Payer responses are still forming: updating medical necessity policies, increasing audits, developing anti-provider AI policies, or building their own AI with platforms like Distyl.

M&A signals the stakes. Arsenal Capital Partners agreed to acquire Knowtion Health, an AI-enabled revenue cycle claim resolution provider, in August 2024. Infinx acquired i3 Verticals' healthcare RCM division in May 2025. FinThrive showcased agentic AI at the HFMA Conference in June 2025. The competitive set includes AGS Health, AdvantEdge, CareCloud, Conifer, McKesson, SSI Group, Athenahealth, Change Healthcare, R1 RCM, GE HealthCare, and Oracle.

AKASA's $85 million-plus backing from Andreessen Horowitz positions it to scale fast in a market where 80% remains untapped. The hiring surge reflects a simple calculation: providers are buying, the workflow layer is wide open, incumbents are reacting, and the window to establish the generative AI standard for revenue cycle operations is now.

Where the Talent Pipeline Breaks

The shortage isn't abstract. It's two distinct labor markets that barely overlap. On one side: revenue cycle professionals who know CPT codes, payer rules, clinical documentation integrity workflows, and the emotional weight of a patient financial conversation. On the other: machine learning engineers who build LLM pipelines, design retrieval-augmented generation systems, and optimize inference latency. Almost no one sits in the middle.

The data makes the gap visible. Thirty-five percent of U.S. medical groups say medical coder is the single hardest revenue cycle role to fill. "Coders aren't cheap and there's not a lot of them," said Jonathan Wiik, vice president of health insights at FinThrive. Nearly half of U.S. hospitals ran negative operating margins in 2022, and rising labor costs have only deepened the squeeze. Retaining specialized staff for CDI, coding, and denials management has become "increasingly challenging and costly," a 2025 Solventum guide found. Meanwhile, 63% of healthcare organizations already use AI and automation in the revenue cycle: 48% apply it to documentation and coding, 73% expect the biggest impact on prior authorizations, and 67% see it driving denials and underpayment recovery. Demand for hybrid fluency is accelerating while the supply side stays frozen.

Why the disconnect? Revenue cycle expertise is earned through years of operational exposure: learning how a specific payer interprets a modifier, when a clinical note supports medical necessity, how to structure an appeal letter that actually gets paid. That knowledge lives in workflows, not textbooks. Generative AI expertise, by contrast, requires fluency in transformer architectures, prompt engineering, evaluation frameworks, and MLOps — skills concentrated in tech companies and research labs, not hospital business offices. The career ladders don't cross. A senior denials analyst doesn't pick up PyTorch on the side; an ML engineer doesn't learn the nuance of a commercial payer's medical policy by reading CMS manuals.

The market friction shows up in adoption barriers. Fifty-one percent of organizations cite IT infrastructure limitations as the top obstacle to AI in the revenue cycle. Forty-four percent lack budget. Forty-three percent struggle with integration. Forty-two percent can't demonstrate ROI. And "securing IT talent is also critical in terms of ensuring organizations have the right in-house knowledge and skills to support AI technologies," the HFMA-FinThrive poll noted. Governance adds another layer: data ownership, privacy, accuracy guarantees, contingency plans for model failure. These aren't problems a pure-play engineer or a pure-play operator can solve alone.

AKASA's hiring pattern reflects that reality. Its 9–10 open roles (mapped on Zero G Talent's board) cluster around the intersection: one, a Senior Forward Deployed Engineer in New York, a Senior TPM in that hub, plus backend, DevOps, and security engineers. The forward-deployed role is telling: it puts engineers directly in customer environments, forcing them to learn revenue cycle workflows by living inside them. The technical program manager sits between product, clinical, and engineering — a translator role that only works if the person speaks both languages.

The company's broader strategy mirrors the "layered" workforce model the HFMA 2026 outlook describes: AI handles rules-based, high-volume processes; a globally integrated workforce manages scale and complexity; governance and analytics ensure accountability; human expertise preserves empathy in financial conversations. AKASA is investing in that integration: standardized training across domestic and offshore teams, shared denial analytics dashboards, real-time escalation pathways, and "AI-enabled workflow orchestration" that routes work to the right layer. It's also funding structured empathy and communication training: scenario-based patient simulations, de-escalation coaching, cultural sensitivity alignment, financial literacy frameworks. The bet is that hybrid talent isn't found — it's built, through deliberate organizational design that forces the two worlds to collide daily.

"The revenue cycle workforce is no longer confined by geography. It is defined by integration," the HFMA report concluded. AKASA's hiring surge is a test of whether a venture-backed AI company can engineer that integration faster than health systems can hire their way out of the same hole.

Four Hiring Lanes — and Where AKASA Fits

Healthcare AI hiring accelerated from a standing start. Through the first four months of 2026, KORE1's healthcare IT desk ran 41 active AI searches against 17 in the same window a year earlier — a 2.4x lift on a desk that does not chase trends. Most U.S. health systems are doubling their AI headcount this calendar year. The payer-and-provider segment, historically the slowest adopter, posted the strongest adoption rise in NVIDIA's 2026 survey: 56 percent of organizations now actively use AI, up from 43 percent in 2024, and 69 percent use generative AI or large language models, up from 54 percent overall.

The market has sorted into three dominant hiring patterns. Ambient clinical AI scribes drove the largest single share, roughly one in three searches KORE1 handled this year tied to an ambient initiative at vendors such as Abridge, Suki, Nabla, or Nuance DAX, or at health systems rolling them out. Abridge closed a $300 million Series E in mid-2025; Suki and Nabla each raised follow-on rounds north of $100 million in the same window. Microsoft's DAX Copilot reported well over a hundred thousand monthly active clinicians by end of 2025. The second pattern is health systems building internal AI governance and platform teams. Chief Medical AI Officer became a defensible standalone C-suite seat in 2026 at most top-fifty U.S. health systems: HCA Healthcare named one in late 2025, Kaiser Permanente in Q1 2026, Providence elevated its seat formally this spring, and Mount Sinai split clinical AI strategy out from the CMIO's office in February. The third pattern sits in biotech and pharma: 48 percent of pharmaceutical and biotechnology respondents use AI agents for drug discovery and biomarker identification, and 46 percent report ROI from AI in drug discovery and development.

AKASA's expansion occupies a fourth, distinct lane: revenue cycle operations automation. Its platform applies generative AI trained on clinical and financial data to claims and prior-authorization workflows. That is not clinical documentation, not drug discovery, and not health-system governance. It is the operational layer where hospitals and payers lose billions to denials and administrative friction. NVIDIA's survey found 85 percent of management respondents say AI increased annual revenue and 80 percent say it decreased costs; 85 percent plan larger AI budgets in 2026, with 47 percent prioritizing workflow and production-cycle optimization.

The talent requirements diverge sharply across these lanes. Ambient scribe vendors need engineers fluent in clinical documentation workflows and real-time audio processing. Health systems need FHIR-fluent ML platform engineers ($190K–$295K) and governance leaders ($320K–$720K base). Biotech needs machine learning scientists with genuine biological domain knowledge, a combination Panda International calls "rare." AKASA needs something different: professionals who understand revenue cycle mechanics (denial codes, payer rules, prior-auth hierarchies) and can apply large language models to them. The board data shows salary bands competitive with ambient scribe vendors but calibrated for an operations domain, not a clinical one.

Geography and work-model expectations are converging. Healthcare AI hiring concentrates in seven metros (Boston, Bay Area, Nashville, Minneapolis, New York, Seattle, Chicago), and remote-eligible roles at health systems dropped from 58 percent of postings in 2024 to 31 percent in April 2026. The default for senior ML hires at large integrated delivery networks is now hybrid, two to three days on campus. AKASA's postings reflect this: senior backend and DevOps roles anchor in both Bay Area hubs; forward-deployed engineering sits in New York City; ML engineering remains remote-eligible across the United States.

Regulatory momentum reinforces the trend. The FDA's traditional device paradigm was not designed for adaptive AI/ML; its January 2025 draft guidance on lifecycle management and marketing submissions for AI-enabled device software functions signals a maturing framework. Meanwhile, 40 percent of large-organization respondents cite compliance with HIPAA, FDA approval, and GDPR as a top factor shaping agentic AI implementation. AKASA's platform (GenAI so trained for revenue cycle) sits squarely in the zone where operational automation meets regulatory scrutiny.

Pure-play AI entrants are arriving from a different vector. OpenAI launched health features in ChatGPT in July 2026, introduced GPT-Rosalind for life sciences research in April, and published work on how agents transform work in June. Their hiring leans toward generalist AI researchers and product teams adapting foundation models for healthcare use cases, not the domain-deep, workflow-embedded profiles that AKASA, ambient scribe vendors, and health systems all require. The bottleneck across every lane remains the same: combinations of skills that almost no one in the market has. KORE1 saw job descriptions asking for five years of ambient AI experience, current FDA SaMD submission experience, fluent FHIR R4 plus SMART on FHIR, prior production deployment at an integrated delivery network, and a master's or PhD in a quantitative field. There are maybe forty people in the United States who satisfy that exact combination. AKASA's screening for hybrid clinical-AI-revenue cycle fluency is a variant of the same structural shortage, just applied to a different, high-value operational surface.

The roles on AKASA's board today will not stay open long. The same workflow fluency that makes a forward-deployed engineer effective at a client site is the currency every lane is bidding for. When those seats fill, the platform thickens. When they don't, the window narrows.

Working in frontier tech? Zero G Talent tracks the openings: see every open AKASA role, browse frontier tech jobs, the companies hiring, and the people building the field.