Working at AKASA: Culture, Pace and Who Thrives

How Work Gets Done at AKASA

Cleveland Clinic began rolling out AKASA's generative AI coding assistant across its U.S. locations in spring 2025, a live production contract that processes more than 100 clinical documents in 90 seconds. That deployment sits at the center of a ~110-person organization whose culture reflects the enterprise, heavily regulated environment its buyers demand: operational intensity organized around compliance-grade precision.

The rhythm doesn't resemble a typical AI startup. No midnight Slack pings chase a model release. No "move fast and break things" mantras adorn the walls. Instead, the pace follows health-system procurement cycles, HIPAA audit trails, and the quiet discipline required when software touches patient revenue. That constraint shapes every standup, every retro, every decision about what ships this sprint.

AKASA operates remote-first with hubs in South San Francisco, New York, and Denver, and team members across 29 states. BuiltIn's September 2026 profile counted 110 employees, 48 in product and engineering. Day-to-day collaboration runs on Slack, Confluence, Jira, and Coda, with cross-team standups and retrospectives baked into the calendar. Leadership publishes OKRs company-wide and treats culture as a company-level OKR, not an HR afterthought. At a 2025 offsite, an entire session tackled silo-breaking, a signal that coordination gets managed explicitly, not left to osmosis.

Decision-making follows the enterprise sales cycle. The four stated values — Empower Health Systems, Innovate for Impact, Win or Learn Quickly, In This Together — appear in bi-annual awards at offsites, a dedicated Slack channel that surfaces both wins and "meaningful learning experiences," and all-hands that recognize tenure alongside group accomplishments. Spot bonuses (cash and equity) arrive ad hoc for exceptional contributions. Formal performance cycles run throughout the year, and nine in ten employees reported in 2025 they feel comfortable sharing challenging feedback, a figure that matters when a missed compliance detail can stall a six-figure deal.

Nearly all employees say they understand how their work connects to company goals. Most feel comfortable asking teammates for help. Most report they can work in a way that fits their life: hybrid, flex hours, caregiving. Manager support scores hit similar highs. The employee Net Promoter Score of 38 lands in the top quartile of tech startups; the engagement score of 89 sits in the top decile. Forbes named AKASA a Best Startup Employer for 2026. Great Place to Work has certified it six straight years.

But the texture matters more than the scores. Work organizes around compliance-grade precision, not because leadership mandates it, but because customers (health systems running revenue-cycle operations) demand it. The pace is deliberate. Accountability is shared. People who thrive treat a delayed launch not as a failure of speed but as a correct calibration of risk.

What Drives the Operating Principles?

AKASA's operating principles read like a direct response to the regulatory and clinical reality of its customers. The company publishes no values page with aspirational slogans. Its principles emerge from how Varun Ganapathi, CTO and co-founder, has described the technical and commercial choices the team made when building the platform, and from the architecture those choices produced.

The foundational principle: human-in-the-loop as a design requirement, not a fallback. Ganapathi put it plainly: "ML enables us to automate the hard stuff. And we believe the special sauce is humans who improve the algorithms by teaching them." In practice, every low-confidence prediction routes to AKASA's own revenue-cycle specialists, who resolve the exception and label the case so the model learns. The loop runs continuously: "when the algorithm isn't sure about what it should do (low confidence), it's escalated to a human-in-the-loop instead. The humans label those examples and identify cases not handled by the current model." This is not a temporary staffing bridge. It is the mechanism by which the system gains the compliance-grade precision health-system buyers demand.

A second principle: deriving automation logic from domains where failure is not an option. Ganapathi said the team "evaluated modern automation approaches from some of the most complex domains in the world (such as self-driving cars), and we derived core principles like the best ways to monitor existing workflows, learn from workflows at scale and quickly adapt to change." The self-driving analogy is deliberate. Both environments require perception of high-variance inputs, long-tail edge cases, and a safety architecture that defaults to human oversight when confidence drops. AKASA translated those principles into proprietary technology "built from the ground up to apply these core principles to the unique challenges of healthcare revenue cycle management."

Third: purpose-built beats retrofitted. Ganapathi has been blunt about robotic process automation: "The hard truth is that RPA is a decades-old technology that is brittle with real limits to its capabilities. It will always have some value in automating work that is simple, discrete, and linear. However, the reason automation efforts often fall short of their aspirations is because life is complex and always changing." AKASA's platform was constructed for the non-linear, exception-heavy workflows of claims, denials, and prior authorizations, work that RPA scripts break on. The result: "over time, the ML model builds resilience to these new edge cases… the automation gets better and better and human intervention will decline over time."

Fourth: deployment speed and remote operability are non-negotiable. The pandemic forced health systems to send revenue-cycle teams home overnight, making traditional consultant-led shadowing impossible. AKASA responded by making Unified Automation "deployable entirely remotely" with its expert-in-the-loop interface handling exceptions in real time. "With continuous machine-learning built-in, the need for costly or time-consuming upgrades and maintenance is eliminated." This principle — remote-first, low-maintenance, self-improving — aligns with the buying cycle of enterprise health systems that cannot tolerate long implementations.

Fifth: the commercial model mirrors the technical model: outcome-linked, not seat-linked. The platform sells on its ability to "accelerate cash collections and decrease the cost-to-collect on every dollar," enabling health systems to "invest more in patient care and be better stewards of the healthcare dollar." Forrester research cited by Ganapathi underscores the economic logic: for every dollar spent on RPA, an additional $3.41 goes to consulting resources. AKASA's architecture eliminates that consulting layer by baking expertise into the platform and its specialist team.

Notably absent from the founder narrative: any mention of speed for speed's sake, hackathon culture, or consumer-AI iteration velocity. The principles that do appear — human-in-the-loop safety, domain-derived rigor, purpose-built architecture, remote deployability, outcome-linked economics — all calibrate to a buyer operating under CMS rules, HIPAA constraints, and audit risk. The culture that follows is one where precision is the currency, patience the timeline, structure the enabler.

The Hiring Bar Selects for Finished Professionals

AKASA's open roles read like a seniority filter set to "experienced." The company's job board lists six recent postings, every one a senior title: Senior Product Manager, Senior Software Engineer (backend and DevOps), Senior Machine Learning Engineer, Senior Technical Program Manager, and a Sales Director. Salary bands cluster between $140,000 and $230,000 with a median of $200,000, per public job postings. That compensation floor alone screens out early-career candidates; the "Sr." prefix does the rest. The message is implicit but unmistakable: AKASA hires people who have already shipped in regulated or complex environments and can operate without hand-holding.

The role mix reflects the company's trajectory. Engineering and ML roles dominate numerically, but the dedicated Sales Director role signals that go-to-market muscle is now a first-class investment. The Sales Director requirement ("remote, United States") suggests AKASA wants quota-carriers who can navigate hospital procurement cycles and security reviews without a local office anchor.

Glassdoor aggregates 26 interview questions and 28 candidate reviews for AKASA (the healthcare AI company, not the Indian airline whose cabin-crew height requirements and group-discussion rounds populate the same search results). That volume indicates a repeatable, documented process, not an ad-hoc conversation.

The South San Francisco office (AKASA's physical hub) adds another signal. BuiltInSF characterizes it as "community-oriented, trust-based" with "connection-building rituals, visible recognition, and high ownership." That description selects for people who show up, speak up, and stay accountable when the org chart is flat enough that ownership isn't assigned — it's claimed. Remote-eligible roles (Senior ML Engineer, Sales Director) still expect candidates to operate within that culture: synchronous collaboration windows, documented decisions, a bias toward over-communication when compliance artifacts are on the line.

Healthcare revenue-cycle automation imposes its own filter. Comparably notes AKASA "scales human intelligence with AI and ML trained on customer data to learn unique systems, adapt to changing environments, and deliver comprehensive automation and analytics for complex workflows." Translate that to hiring: the company needs engineers who understand that "customer data" means protected health information, that "changing environments" means payer policy shifts and CMS rule updates, and that "complex workflows" means prior-authorization logic trees that vary by provider, specialty, and region.

The Sales Director role makes the enterprise filter explicit. Selling into health systems means multi-stakeholder deals (CFO, CIO, CMIO, revenue-cycle VP), pilot-to-contract cycles measured in quarters, and procurement processes that treat AI as a risk category until proven otherwise. The hiring bar selects for candidates who have closed six-figure SaaS deals into hospitals or large physician groups, who know the difference between a Business Associate and a Business Associate Agreement, and who can articulate ROI in denial-reduction dollars — not "AI-powered efficiency."

What's absent from the public record is equally telling. No entry-level roles. No "growth" or "potential" language in the postings. No mention of rotational programs, mentorship structures, or tuition reimbursement. AKASA's hiring bar selects for finished professionals who can contribute immediately to a ~110-person organization operating in a regulated domain where mistakes cost money and credibility. The trade-off: higher autonomy, higher compensation, higher stakes. Candidates who need structure, mentorship, or a consumer-product pace will self-select out — or be selected out — before the first technical screen.

The Employee Record: Two Pictures, One Workforce

On Blind, half a dozen reviews yield a two-and-a-half out of five average. Management scores 1.7. Career growth sits at 2.0. Work-life balance and culture each hit 2.5. Only compensation and benefits reaches 2.8. By contrast, AKASA's own BuiltIn profile (updated September 2026) claims nine in ten employees call it a great place to work, an engagement score of 89 (top decile for tech startups), an eNPS of 38 (top quartile), and six consecutive Great Place to Work certifications. The gap between the two sources is the story.

Praise, sourced and dated

Compensation and benefits draw the most consistent positive notes on Blind. A February 2025 reviewer wrote: "Pay and benefits are extremely competitive. Company is fully remote." Another from the same month: "Most people were good people, benefits were great." An April 2022 review highlighted "interesting field of ML in healthcare. Great problems to solve and market is big." A May 2022 review listed "unlimited PTO; manager always approve your PTO" and "lots of free t-shirts." A June 2023 reviewer noted "good wlb. People are generally nice at IC level, management is good."

The BuiltIn page (2026) surfaces three named testimonials. One describes "a profound sense of purpose and fulfillment… our innovations directly contribute to enhancing the overall patient care experience" and says they've "grown immensely as an engineer." Another calls out "collaborative and supportive culture, meaningful and challenging work, flexibility, strong leadership, and competitive pay and benefits" as themes employees raise. A third states: "Many companies get one of those right… but it's exceedingly rare that a company gets both right. AKASA is one of those places." The page also lists published perks: unlimited PTO, robust health plans with some 100-percent-cost tiers, Spring Health, One Medical, parental leave, 401(k), and funded off-sites.

Criticism, sourced and dated

Management and direction dominate the negative Blind feedback. The February 2025 review that praised pay also said: "Took years and millions of wasted dollars before senior leadership pivoted toward things AI/ML could actually provide value for in the target market." The same reviewer added: "Hard to advance your career. Poor leadership. Teams were underresourced." A June 2023 reviewer cited "uncertain future (normal for startups) kind of normal startup bs, tc lower, somewhat chaotic upper level directives." Two May 2022 reviews are harsher: "Lots of opportunities because many decisions were made quite poorly" and "Put out fires for days. Managements change rules often before you can get jobs done." An April 2022 review called out "very inexperienced ML people and don't know what is going on. No clarity on the vision and poor execution. Wouldn't recommend people joining this unmanaged ship." A May 2022 entry noted "benefits are not bad but compensation is terrible… very complex promotion system that's more suited for mega cap companies… management is willing to hear our feedback but doesn't act on most of them."

Career growth scores 2.0 on Blind, and multiple reviews flag a promotion process built for a much larger organization. The BuiltIn page counters with a documented internal-promotion policy, quarterly engagement surveys, and a claim that employee feedback shapes policies, but the Blind reviewers writing in 2022–2025 describe the opposite experience.

Who Thrives and Who Burns Out

The split in AKASA's employee record is not subtle. Both data sets are real. The gap between them is the culture.

People who thrive here tend to share three traits. First, they draw energy from the mission: deploying generative AI into the mid-revenue cycle of major health systems, like the Cleveland Clinic deployment noted earlier, where a model reads a clinical document in under two seconds and processes more than 100 documents in 90 seconds. That work is tangible. One engineer on BuiltIn wrote they'd "grown immensely" because they're "constantly presented with opportunities to expand my technical expertise and problem-solving skills." Another echoed the profound sense of purpose and fulfillment noted in the testimonials above, knowing that their innovations directly enhance the overall patient care experience. If you need that kind of impact signal to stay motivated, AKASA delivers it.

Second, they function well in a fully remote, async-first environment with unlimited PTO that managers actually approve. The company lists flexible work schedules, hybrid options, and defined in-office expectations as formal policies. Multiple Blind reviewers from 2022 to 2025 flagged "good wlb" and "manager always approve your PTO" as positives. For individual contributors who value autonomy over oversight, the structure holds.

Third, they can operate inside a compliance-grade sales motion without waiting for permission. The first-party board data shows open requisitions for a Sales Director at $150k–$185k alongside senior engineering and ML roles at $175k–$240k. That ratio shapes daily work: enterprise deals, regulatory review, and documentation rigor set the tempo. Candidates who have sold into hospitals or navigated HIPAA, SOC 2, or payer-specific requirements will recognize the cadence. Those who haven't will learn it on the job.

Who struggles? The Blind record is consistent across years, echoing the criticisms detailed earlier: management not acting on feedback, constant firefighting, chaotic directives, limited career advancement, wasted resources before strategic pivot, inexperienced ML talent, unclear vision, and a promotion system ill-suited for the company's size.

The burnout profile is clear: you want clear direction, fast promotion, stable leadership, and a mature engineering culture. AKASA offers none of those consistently. What it offers is a live production contract with Cleveland Clinic, a revenue-cycle domain that isn't going away, and a remote-first workplace where the mission is specific enough to keep some people for years while others leave once they realize the org chart, the promotion ladder, and the technical strategy are all still being written — calibrated to the speed of a health-system procurement cycle, not a sprint demo.

Our investors

Andreessen Horowitz, according to AKASA's careers page, is a Silicon Valley-based venture capital firm with over $10B in assets under management.

BOND is a global technology investment firm that supports visionary founders throughout their entire life cycle of innovation and growth.

Costanoa Ventures is a Silicon Valley-based venture capital firm which invests in enterprise-focused seed and Series A technology companies.

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