1.2 Million Calls Placed by Careforce AI Lift Renewals to 96%

Five Open Roles, One Salary Band

Careforce, a Y Combinator-backed startup building autonomous AI agents that call, schedule, and coordinate patient care, has five open positions on its hiring board as of October 2026. Nine employees work from San Francisco and Oakland. Their agents already handle outreach in 29 languages using more than 50 scripted conversations, Y Combinator's healthcare startups directory shows 50+ outreach scripts, closing four in five calls into booked appointments, Y Combinator's healthcare startups directory reported an 80% conversation-to-booking closing rate. That traction, paired with a reported 6x ROI for healthcare organizations, drives the hiring push, and a screening process that now serves as the primary gatekeeper for every candidate.

The roles span sales, delivery, engineering, and product. Four of five carry a base salary band of $140,000 to $180,000; the Customer Success & Delivery Lead lists $120,000 to $140,000. Locations cluster in the East Bay and San Francisco, with several listed as remote within the U.S.

Enterprise Sales Lead owns the full cycle: pipeline generation, demo, negotiation, close. The posting emphasizes healthcare buyers (health systems, payers, value-based care groups) who need to see ROI before they sign. Technical Program Delivery Lead sits between sales and engineering. Once a deal closes, this person scopes implementation, coordinates the customer's IT and clinical stakeholders, and ensures the AI agents go live without integration work on the client side. Careforce's agents "require no integration" and "navigate all tools," per the company's description, so the delivery lead orchestrates rather than custom-builds.

Product Engineer on the Care Loop team builds the core agent runtime: the orchestration layer that lets an LLM drive phone, text, email, and mail outreach; the memory system that tracks patient context across 50-plus script branches; the evaluation harness that measures conversation-to-booking rate. The role requires production-grade Python, experience with LLM tooling, and comfort shipping to a regulated environment. Forward Deployed Engineer is the field counterpart. They embed with early customers, instrument the agent's behavior in live workflows, and feed failure modes back to the product team. The posting calls for "make agents more capable with interfaces to the real world," a direct quote from Careforce's site, which in practice means writing the glue code that lets an agent navigate an EHR, a scheduling system, or a fax portal without an API.

Customer Success & Delivery Lead, posted October 6, 2026, owns the post-launch relationship. The mandate: "make every customer feel like our only customer. Except our customers' frontline workers are fully autonomous AI agents, calling." That line captures the hybrid nature of the role — part account management, part operations. The lead monitors agent performance dashboards, runs quarterly business reviews with clinical leadership, and coordinates retraining when a new outreach script or language is added.

A sixth role, Technical Product Manager in San Francisco, appears on LinkedIn but not on the Y Combinator jobs board as of the latest scrape. The board data takes precedence; the five roles above are the ones currently live on Careforce's primary hiring channel.

All five positions share a common thread: they require operating at the boundary where an autonomous agent meets a messy, regulated healthcare workflow. The salary bands are tight. The roles do, with remote options as noted. The screen is what filters for people who can actually ship in that environment.

What the Screen Actually Tests

Careforce's job postings read less like wish lists and more like a map of the exact problems the nine-person team faces today. The company is developing such agents that call patients in 29 languages, navigate fragmented portal ecosystems, and close care gaps for safety-net organizations like Kern Family Health Care, where agents have already placed more than 1.2 million calls, over half in Spanish, and helped lift renewal rates from 38 percent to 96 percent. That operational reality shapes every screening criterion.

Healthcare Fluency Is Non-Negotiable

Every open role lists healthcare industry insight as either a requirement or a "strong plus." For the Technical Product Manager role, the posting explicitly calls for "experience navigating healthcare administration and a deep understanding of provider and patient pain points." The Customer Success Delivery Lead description goes further: familiarity with payers, providers, care coordination, and comfort working with PHI under HIPAA. A third-party source notes the platform was "designed to support workflows that often live across fragmented data systems and vary from clinic to clinic," meaning candidates who have never touched an Epic, HL7, or FHIR integration will struggle to contextualize the product they're selling, supporting, or shaping.

Founder Sial's background, calculating risk scores for health plans before starting Careforce, set the template. The CHCF profile quotes him: "Health plans use math and science to figure out how to provide care, but the problem is that many people never even come in to receive care." That framing — access, not just efficiency, is the lens through which every hire is evaluated.

Agent-Native Technical Chops

"Building agentic systems is still a nascent skill," the careers page warns. "It is easy to end up getting into old habits and building SaaS." The Customer Success Delivery Lead role demands hands-on LLM/agent experience: "you've worked with LLM-powered products or AI agents (voice or chat), and you understand prompts, tools, evaluations, and why agents fail." The Technical Product Manager posting requires "a solid grasp of AI technologies and their applications" and the ability to "engage engineering teams on technical concepts."

This isn't abstract AI literacy. Careforce's agents handle real-time voice conversations, escalate to humans when stumped, and integrate with portal infrastructure across five counties. Candidates who have only built RAG chatbots or prompt-tuned classifiers will hit a ceiling fast.

Data-Driven Product Instincts

Both posted roles emphasize analytical mindset. The PM role: "you use data to drive product decisions and have experience setting and evaluating product metrics." The CS role: "you use data and clear metrics to track customer health and drive decisions." In practice, this means comfort with SQL (listed as "a plus" for CS), reading logs, querying production databases, not just dashboard gazing. The company's own impact metrics (3x appointments scheduled, 80% care gaps closed, 96% renewal rate) are the language the team speaks internally.

Startup Velocity and Ownership

"Adaptability and initiative: you thrive in a fast-paced startup and take a proactive approach to overcoming challenges." "Ownership and urgency: you treat a customer issue as yours until it's resolved, and you also thrive in such an environment." These phrases appear nearly verbatim across both postings. With a team of nine, there is no specialized ops, devops, or QA function to fall back on. The CHCF article notes integration timelines of "several weeks" where traditional health-tech rollouts take months, a pace that demands engineers and customer-facing staff who can ship a prompt fix, update a workflow config, and brief a clinic director in the same afternoon.

The Pattern Behind the Postings

Synthesize the two live roles and a clear profile emerges: Careforce screens for people who have operated at the intersection of healthcare operations, LLM-backed automation, and early-stage product velocity. The screen filters for translational fluency — the ability to move between a clinic's fax-dependent referral loop and a prompt chain that schedules the colonoscopy, because that is the only way the product works.

Inside the Interview Loop

Careforce moves fast once a candidate clears the initial filter. The interview process spans three stages over roughly seven days, faster than the ten-day median for comparable healthcare and healthcare IT startups, and it skips the take-home assignment entirely. Glassdoor users rated their interview experience at Careforce as 77.8% positive with a difficulty rating score of 2.44 out of 5. The Anti Job Board models the process at three stages for a <50-person Healthcare, Healthcare IT company; treat the following as a model, not confirmed detail.

Stage 1: Intro Call: 30 Minutes, Video

The first conversation is a culture-fit and expectations check, typically led by a founder or the hiring manager. At this size, "hiring manager" effectively means a founder who also owns the product roadmap and the sales pipeline. The call tests whether the candidate understands what Careforce is building — AI workers that find, call, and schedule patients for healthcare organizations, and whether they can operate without close supervision. Autonomy is weighted heavier than algorithmic depth; research shows that showing you can work independently "matters more than algorithms at <50 people."

Stage 2: Technical Deep Dive: 60 Minutes, Video or In-Person

The second round is a working session with the technical founder or a lead engineer. It runs an hour and focuses on past projects and problem-solving approach rather than whiteboard puzzles. Glassdoor reviews (nine interviews, nine reviews as of the latest scrape) describe the tone as collaborative. Candidates walk through a system they built, the trade-offs they made, and what they would change in hindsight. Because Careforce embeds autonomous agents into clinical workflows, the conversation often drifts into reliability, observability, and failure modes. The absence of a take-home means this round carries the full technical evaluation weight.

Stage 3: Final Round: 45 Minutes, In-Person or Video

The last stage brings in the broader founding team. It tests team fit and doubles as the offer discussion. The "most challenging round" label attached to this stage in the research likely stems from the breadth of evaluation: each founder probes a different dimension: product intuition, clinical workflow awareness, communication style with non-technical stakeholders. At nine people, a hire changes the team's dynamic materially, so the conversation is less about checking boxes and more about whether the candidate's default operating mode aligns with a group that ships fast, talks to customers daily, and has no dedicated QA, DevOps, or people-ops functions.

Competitive Context

Fifty to one hundred applicants typically apply within the first two weeks for roles at this stage and sector. The seven-day timeline means the process compresses quickly, a candidate who passes the screen on Monday can have an offer by the following week. That speed cuts both ways: little room to "circle back" or request additional time to prepare. A day-five follow-up can double response rates in this segment, but once the interview clock starts, the cadence is set by the founders' availability.

How to Prepare

The research available on Careforce's hiring process is notably thin. Unlike larger AI labs that publish interview guides or have candidates openly debrief on forums, Careforce, a team of nine per Y Combinator and its own listings, has not left a detailed public trail of its interview stages, rubric, or candidate feedback. What we know comes from its product footprint: a voice agent (Angelica) and a text agent (David) built on Twilio Voice and Messaging, deployed in HIPAA-compliant environments across safety-net health plans and FQHCs in five California counties, with CHCF's $500,000 investment and Stella Tran of the CHCF Innovation Fund calling it "one of the first companies building AI agents for safety net organizations." That context shapes what a credible candidate can prepare for, even without a leaked interview playbook.

First, understand the technical stack and the constraints it implies. Careforce's agents operate in 29 languages, hand off to human staff when confidence drops, and integrate into fragmented EHR and eligibility systems "in several weeks" rather than months. Candidates should be ready to discuss: building voice-first LLM applications with strict latency budgets (Twilio's Phonely benchmark cites 99.2% accuracy at ultra-low latency); designing guardrails for HIPAA-covered conversations where a hallucination triggers a state grievance; writing evals for multi-turn dialogues that span eligibility verification, appointment scheduling, and transportation coordination. If you've shipped a voice agent that passed a SOC 2 audit or integrated with an Epic or eClinicalWorks instance, lead with that. If you haven't, show you've worked with Twilio's Voice SDK, WebRTC, or comparable telephony APIs and can speak to the failure modes — background noise, accent variation, mid-call interruptions, that don't appear in text-only benchmarks.

Second, demonstrate fluency in the operational reality of safety-net healthcare. The KFHC case study is the clearest window: 270,000 members, auto-renewal ending July 1, 2025, a jump from 38% to 96% renewal rates, and 1.2 million calls with over half in Spanish. Candidates who can articulate why "three months before expiration" beats "one month before" — because staff bandwidth is the bottleneck, not model quality, signal they've thought beyond the model layer. Read the CHCF blog post (July 16, 2026) and be ready to discuss how Angelica's language detection and handoff logic reduce grievance risk. That's product sense, not just engineering.

Third, expect the screen to test whether you can operate in a "clinical operations" mindset, not just an "AI research" mindset. The company's own description — "autonomous AI workers that schedule and manage healthcare appointments", positions it closer to an AI-native BPO than a model lab.

Fourth, prepare for a practical assessment that mirrors the actual work. While no public rubric exists for Careforce, peer companies in the Twilio Searchlight cohort (Insight Health AI, OhMD, Phonely, Strada) commonly use voice-agent flow exercises, live debugging sessions, and call-transcript reviews. Treat any take-home as a production prototype — include logging, fallback prompts, a handoff path, not a notebook demo. Note: the Anti Job Board model states Careforce does not include a take-home stage.

Fifth, leverage the network that exists. CHCF's portfolio, Y Combinator's healthcare batch (277 companies), and Twilio's Searchlight alumni are dense with people who've either worked with Careforce or evaluated similar vendors. A warm introduction from a founder whose company integrates with Twilio Voice for healthcare, or a clinician who's used Angelica at KFHC or Ritter Center, carries more weight than a cold application. The CHCF blog post names Alonso Hurtado (senior director of corporate services at KFHC) and Cesar Delgado (CIO at KFHC), their teams are the users. If you can reference a conversation with someone on the receiving end of Careforce's calls, you've done homework most candidates skip.

Finally, be honest about the stage. Careforce is post-pilot, pre-Series A, and hiring into a problem space where the next federal rule change (H.R. 1, effective January 2027, requiring 80 hours/month documentation for Medicaid eligibility) will multiply demand for exactly this automation. The interview is as much about whether you want to build in this regulatory-operational crucible as whether you can pass a coding challenge. Candidates who frame their motivation around "reducing administrative burden on safety-net staff so patients don't lose coverage" — Sial's stated thesis, align with the mission the board and investors backed. Those who treat it as a generic voice-AI role rarely make it past the screen.

The screen doesn't care about your pedigree. It cares whether you've ever made an agent work inside a clinic that still runs on fax.

Working in AI? Zero G Talent tracks the openings: see every open Databricks role, browse AI jobs, openings at Anthropic, and the people building the field.