Principal AI-Native Product Builder

Cogitativo; United States; On-site; USD 180,000 - 280,000; Full time

Job summary

Cogitativo in Berkeley seeks a Principal AI-Native Product Builder to own the product from concept through launch and beyond, blending product strategy, UX design, and AI-assisted development into one fluid workflow. You will define problem-first roadmaps, validate opportunities with evidence, and ship production-ready software using autonomous building loops, powered by generative AI and AI-native tools.

Healthcare domain knowledge and the ability to drive impact across payer and provider

Qualifications

• You start with the customer challenge, not the tech, and can reframe a problem or decide not to build.
• A track record of shipping software independently that real people used.
• 10+ years in product management, product design, software engineering, or founding roles with 5+ years owning outcomes for a shipped product.
• Design skills to create usable interfaces without a designer and impress with a proven demo.
• Deep fluency in LLM capabilities, context windows, and prompt architectures with AI tooling awareness.
• Solid understanding of API design, data stores, auth flows, and deployment infra.

Responsibilities

• Understand the customer problem deeply and size the opportunity.
• Establish value to the customer and to Cogitativo; define the business rationale and cost to build.
• Decide what AI to build vs. not build and guide roadmap with evidence.
• Design AI-first UX and ownership of product metrics (adoption, retention, unit economics).
• Own deploy, observability, and telemetry; ensure tests and rollback plans are in place.

Skills

Problem-first thinking

Builder mindset

Product judgment

Design taste

AI literacy

Technical fundamentals

Communication

Healthcare domain knowledge

Managed care economics

Claims and payment integrity

Government payers

Clinical and population health

Healthcare data

Regulatory and privacy

Tools

API design

SQL/NoSQL

Auth flows

Deployment infrastructure

Vercel

Supabase

AWS

Job description

Cogitativo is a Berkeley-based leader in healthcare AI and data science, bringing together nationally recognized physicians, clinical researchers, healthcare operators, and data scientists to tackle the most complex challenges in healthcare. Our three integrated platforms — Visión (precision health), Conciencia (operational intelligence), and Carta (delivery economics: causal precision on cost, contracting, policy, and network decisions) — deliver AI-driven capabilities with measurable, real-world impact.

We are seeking a Principal AI-Native Product Builder who will own the product they build and set the pattern for how Cogitativo builds software in the age of agentic AI.

Role overview

You will own the product you build, from concept through launch and beyond: the problem, the design, the code, the models inside it, and the metrics after launch. You merge product strategy, UX design, and AI-assisted development into one fluid workflow.

Everything starts with the problem. Before you build, you dig in: who the customer and user are, what is hard about their day, why existing tools fail them, and why Cogitativo is the right company to solve it. Speed only matters when pointed at a problem where we can make an instrumental impact.

Instead of writing PRDs for a separate engineering team, you use natural language, AI IDEs, and agentic workflows to build, iterate, and deploy production-ready software yourself. With generative AI as your leverage, you deliver what a traditional pod (PM, designer, two engineers, QA) would deliver, at a fraction of the cycle time.

This is a Principal-level individual contributor role.

Why this role exists

The cost of turning an idea into working software has collapsed. The bottleneck has moved from writing code to deciding what is worth building and judging whether what got built is good. The PM-to-designer-to-engineer-to-QA handoff was built for a world where each step took weeks.

We are hiring one person to collapse that loop: customer problem to working software in days, not quarters. New capabilities in a day, a whole product in a week, telemetry acted on as soon as it lands.

Key responsibilities

Everything starts with the business rationale and product strategy: understand the customer problem deeply, size the opportunity, decide whether it is worth solving, and make the case for why Cogitativo should tackle it.

Product strategy and design taste

• Problem discovery: sit with care managers, claims examiners, and payer executives; map their workflow; find where the real friction, cost, or risk sits. Validate with evidence, not assumptions.
• Business rationale: for each opportunity, establish value to the customer, value to Cogitativo, and cost to build and run. Tie it to our data, clinical and managed care expertise, customer relationships, and platforms. Be clear when someone else should build it.
• Curation and judgment: separate what AI can build from what we should build. Kill ideas early and cheaply.
• AI-first UX: design adaptive interfaces where the system anticipates user intent, not rigid form-and-button layouts.
• Outcome ownership: define your product's metrics (adoption, retention, task success, unit economics) and move them.

AI-assisted development

• Prompt to production: build, test, and deploy web and mobile frontends and backends with AI-native tools (Cursor, Claude Code, Lovable, v0, and whatever replaces them next quarter).
• Architecture curation: guide AI to write clean, modular, testable code; debug, refactor, or hand-edit when the model hits a wall.
• Rapid iteration: turn user feedback into working changes in hours. Ship small, ship often, measure.
• Production ownership: own deploy, observability, and on‑call hygiene. Nothing ships without tests, rollback, and telemetry.

AI systems and context engineering

• Core AI integration: embed LLM orchestration, RAG, tool‑calling, and agentic loops into the core workflow, not bolted on as a chat box.
• Context and prompt engineering: craft prompts, context structures, and system instructions that yield deterministic, safe, high‑quality outputs.
• Evaluation and latency tuning: build eval frameworks (golden sets, LLM‑as‑judge, regression suites); manage tokens; optimize inference cost and latency.
• Model strategy: choose between frontier model, small model, cached result, or plain heuristic, and revisit as price and capability shift.

Principal-level influence

• Set the standard: tooling, workflows, guardrails, and code‑review norms for AI‑native building across the product organization.
• Multiply others: coach the team and publish playbooks.
• Advise leadership: bring a hands‑on view of what current models can and cannot do into roadmap and investment decisions.

Required qualifications

• Problem-first thinking: you start with the customer challenge, not the technology, and can show times you reframed a problem, made the case for solving it, or decided not to build.
• Builder mindset: a track record of shipping software independently (side projects and MVPs count) that real people used.
• Product judgment: 10+ years in product management, product design, software engineering, or founding roles, including substantial healthcare experience, with 5+ years owning outcomes for a shipped product. You have driven strategy, not just executed it.
• Design taste: you can design a usable interface without a designer and know a demo that impresses from a flow that works.
• AI literacy: deep, current fluency in LLM capabilities and limits, context windows, structured outputs, tool use, developer tools, and prompt architectures. You have opinions about when agents work and when they fail.
• Technical fundamentals: not a staff engineer, but you understand API design and integration, SQL/NoSQL databases, auth flows, and basic deployment infrastructure (Vercel, Supabase, AWS). You can read a stack trace and a schema.
• Communication: you write clearly, demo well, and can explain an AI system to an executive, a customer, and an engineer in their own terms.
• Healthcare domain knowledge (required): our products live inside payer and health system operations. You bring working knowledge of several of these and can learn the rest quickly:
• Managed care economics: cost of care, provider contracting, utilization management, benefit design, payer reporting tools.
• Claims and payment integrity: claims lifecycle and adjudication, 837/835, CMS-1500/UB-04, appeals, fraud, waste, and abuse analytics.
• Government payers: VA (Community Care, VA Financial Management, TeleCritical Care), CMS, Medicare Advantage, Medicaid managed care, DHA; federal procurement and compliance a plus.
• Clinical and population health: chronic disease (ASCVD, CKD, Type 2 diabetes, COPD), risk stratification, care management workflows, HEDIS and Star ratings, SDOH.
• Healthcare data: EHR, claims, pharmacy, lab, SDOH; ICD-10, CPT/HCPCS, NPI; data quality and standard data models.
• Regulatory and privacy: HIPAA and PHI, FedRAMP, state filings (e.g., DMHC), and explainability and auditability for AI in healthcare decisions.

Preferred qualifications

• Products built for the VA, CMS, or large commercial payers, or AI features that cleared a healthcare compliance or security review.
• Hands‑on eval tooling, LLM observability, or fine‑tuning and distillation. li
• A 0-to-1 product launch inside a larger company or at an early‑stage company.
• Security, privacy, and compliance constraints for AI features (PII, data residency, model‑output safety).
• A public body of work: open source, published apps, writing, or talks on building with AI.

What success looks like in 90 days

• Day 30: understanding of Cogitativo's place in the market, understanding of our capabilities, onboarded onto our AI developer environment; 3 to 5 incremental features shipped to production
• Day 60: investigated a customer problem firsthand, written a one‑page case for why Cogitativo should solve it, and built and validated a new internal tool or micro‑feature from scratch using autonomous building loops; in daily use by at least one team.
• Day 90: scaled a net‑new AI‑native product flow to early users; eval suite in CI; a cost‑per‑task number and a plan to cut it; first draft of the team's AI‑native build playbook.

How we evaluate candidates

We weight what you have built over what you have written.

• Take‑home build and walkthrough: we give you a small, real backlog problem as a one‑paragraph brief. Take it to a working, deployed prototype with your own tools, then show us how you framed the problem, directed the model, stepped in, and what you cut.
• Product and judgment conversation: a problem you dug into deeply, how you decided it was worth solving and why your company was the right one, what you chose not to build, and how you weigh AI capability against cost.
• Systems deep dive: an AI feature you built — prompts, context strategy, evals, failure modes, and what you would change now.
• Leadership conversation: scope, influence, and how you will raise the bar for others.

If you have shipped anything you can and want to show us, send links with a paragraph each on the problem, what you used, and what happened. Side projects count.

This role is probably not for you if:

• You prefer to specify work and hand it to others rather than build it yourself.
• You want a dedicated engineering team, a designer, and a two‑week sprint cadence.
• Your AI experience is limited to chat assistants and you have not created new solutions using AI.
• You are uncomfortable owning production incidents for code an AI helped write.
• You need a stable toolchain; ours will change every few months, on purpose.

Cogitativo is an equal opportunity employer. All qualified applicants receive consideration without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, age, disability, genetic information, veteran status, or any other protected characteristic.