• Build agent harnesses in Python using LangChain and LangGraph, including tool-calling, structured outputs (Pydantic/JSON schema), retries, streaming, and memory • Package agent harnesses for the AgentCore Runtime with appropriate context, tools, skills, and subagents that fit cleanly into production flows and scenarios • Write the tools and skills agents use: API integrations, SQL queries against Snowflake, Snowflake backed knowledge retrieval with clear contracts and Pydantic validation • Build evaluation harnesses (golden datasets, LLM-as-judge, regression suites) using AgentCore Evaluations, and wire them into CI • Implement guardrails around tool execution: auth scoping, input/output validation, PII and prompt-injection protections, and hallucination mitigation • You own what you ship: prototype, deploy through Amazon AgentCore, monitor traces, and fix it when it breaks • Partner with data engineers on Snowflake backed retrieval patterns (Cortex Analyst and Cortex Search Services) • Contribute to refining our internal engineering patterns as the stack evolves

1 in 4 people in the US have a treatable mental health condition, but most providers don't accept insurance, making therapy too expensive for most people. Headway’s mission is to fix this by building a new mental healthcare system everyone can access. We started by solving the biggest barrier to care: insurance. The admin work - credentialing, claims, payment reconciliation - is a nightmare. We've automated that. But we're going further. Over 75,000 providers across all 50 states run their practice on our software, serving over 1 million patients. We are building the best tools for therapists to run their entire practice, reimagining the experience of finding a therapist, and investing in the platform foundations to enable this at scale. We aren't just a billing layer; we are becoming the platform where care actually happens. We're a Series D company with $325M+ in funding (a16z, Accel, Spark Capital, etc.), looking for exceptional people to help us achieve this mission. We want your time here to be the most meaningful experience of your career. Join us, and help change mental healthcare for the better. About This RoleYou will be Headway's first AI Platform product manager. Your mission: build the AI platform that makes every Headwegian an AI-native employee. In Q1 2026, we launched Eddy — an internal AI assistant with direct access to Snowflake, Jira, Slack, GitHub, Google Workspace, and more. Early adoption is strong in R&D, but usage drops off sharply outside of it. Eddy is the first of what we believe will be many: an ecosystem of agents, skills, evals, and data integrations that together become Headway's AI platform layer. You'll own that vision — and the roadmap to get there. This is not a classic internal tools role. You are a product builder with deep AI fluency who will ship fast, write evals, configure agents, and sit with CX, Ops, Sales, Finance, and Clinical teams to find the workflows where AI actually moves the needle. Today, three engineers are building Eddy without a PM. You'll be the first, building the vision that the team is currently shipping without. You Will - Set & build the AI platform vision. Define what Headway’s agent harness should look like — from platform & data architecture to the connected ecosystem of agents, skills, and integrations that will sit on top of it. - Own Eddy's roadmap. Replace real-time feature triage with a first-principles product strategy. Build the user research pipeline, the eval infrastructure, and the prioritization framework that don't exist today. - Enable adoption across functions. Partner with functional leaders in CX, Ops, Sales, Finance, Clinical, and more to identify high-value AI workflows, build them, and turn non-adopters into power users. Who You Are - 8+ years of product experience, with at least 2 years shipping AI or AI-adjacent products. - You have personally built or configured AI agents, skills, or workflows — not just managed the engineers who did. You write prompts, design evals, and know where LLMs break. - Strong AI fluency: prompt engineering, context management, RAG, tool use, evals. You can hold your own with ML engineers and have real opinions on cost/latency/quality tradeoffs. - A platform thinker who can build for other builders, not only end users. - A cross-functional operator who translates AI capabilities into business value for non-technical stakeholders — and knows how to say "no" to low-value requests without losing the relationship. - Ruthlessly prioritized. You can hold a product roadmap, a platform roadmap, and an adoption strategy in your head at once and know which to pull on this week. - Above all, you believe AI is going to fundamentally change how knowledge workers operate — and you want to prove it at a company where the CEO agrees with you. We believe a team's strength is in its people, and we cannot achieve this mission without a team that reflects the diversity of this problem – across race, ethnicity, gender, sexuality, age, national origin, religion, family status, disability, military status, and experience. Headway is committed to the full inclusion of all qualified individuals. As part of this commitment, Headway will ensure that persons with disabilities are provided with reasonable accommodations. If reasonable accommodation is needed to participate in the job application or interview process, to perform essential job functions, and/or receive other benefits and privileges of employment, please inform the recruiter when they contact you to schedule your interview. Headway participates in E-Verify. To learn more, click here. A notice to Headway applicants: To protect yourself against phishing and recruitment fraud, please note that Headway only accepts applications through our official careers page at https://headway.co/careers. Headway will never refer you to external websites, ask for payment or personal information, or conduct interviews via messaging apps. All official communication will come from a @findheadway.com email address. If you are contacted by someone claiming to be from Headway via an unofficial channel, please do not share any information and report it as spam.

• Atuar de forma estratégica na nossa estrutura de Plataforma de Engenharia IA. • Referência técnica para a evolução de APIs, SDKs e serviços críticos de inteligência artificial na Vivo. • Definir a visão técnica e os padrões de desenvolvimento para APIs, SDKs e serviços de plataforma em IA. • Liderar a arquitetura de soluções complexas, garantindo que a orquestração de agentes e serviços de IA seja escalável e resiliente em nível corporativo. • Atuar como advisor técnico para lideranças e squads. • Evangelizar boas práticas de engenharia e cultura de qualidade em toda a diretoria de Dados & IA. • Desenhar e otimizar a governança técnica. • Identificar e mitigar riscos técnicos estruturais antes que impactem a operação ou o roadmap de produtos de IA.

Role Description We are looking for a Senior AI Engineer to help build the core intelligence layer behind TaxGPT. This role is focused on improving how effectively, reliably, and safely AI can automate real tax workflows using LLMs, retrieval systems, structured context, and agentic workflows. You will work closely with software engineers, product teams, and technical leadership to design, build and scale AI systems that support fast development, stable production environments, and long-term scalability. This is a senior individual contributor role for someone who can work with high autonomy, move quickly from prototype to production, and bring strong judgment to applied AI product development. What You'll Do - AI Systems and Agentic Workflows - Build and improve AI-powered product experiences for tax and accounting workflows - Increase automation rates across workflows while maintaining accuracy, reliability, and a strong user experience - Design and refine prompt-based systems, tool-using agents, and multi-step workflows that perform well in real production use cases - Prototype and ship end-to-end AI features quickly, from idea to working product - Design workflows that combine LLMs, tools, APIs, retrieval systems, and structured outputs in a robust way - Work across APIs, tools, retrieval layers, and backend systems to make AI capabilities useful in real user workflows - Model Quality, Evals, and AI Reliability - Build and maintain evals, regression tests, and benchmarks that help us measure and improve AI quality over time - Define practical metrics for usefulness, accuracy, latency, reliability, and cost - Investigate model failures and systematically improve performance through better prompting, context design, routing, and system architecture - Contribute to fine-tuning experiments, benchmark design, and dataset development where it adds product value - Applied Engineering & Technical Leadership - Write strong Python code for prototypes, internal tooling, backend services, and AI workflows - Work with product and backend engineers to productionize AI systems cleanly and safely - Use data and experimentation to guide decisions, validate improvements, and prioritize the highest-impact work - Help build feedback loops that make AI behavior easier to understand, debug, and improve over time - Make strong technical decisions in your area and help the team balance speed, quality, and maintainability - Share clear guidance through code reviews, design discussions, and documentation Qualifications - 5+ years of experience in software engineering, machine learning engineering, AI engineering or a related role - Strong experience building with LLMs in real products, not just experimentation environments - Experience designing and improving prompt-based systems to achieve reliable, high-quality outcomes - Strong experience with agentic workflows, tool-using AI systems, or multi-step AI orchestration - Experience building evals, regression tests, benchmarks, and monitoring for AI workflows - Strong understanding of tradeoffs in model quality, latency, cost, reliability, and system design - Strong scripting or coding ability in languages such as Python, Go, or TypeScript - Comfort working with APIs, backend systems, data flows, and cloud environments - Strong written and verbal communication skills Preferred Qualifications - Experience supporting fast-moving startup engineering teams - Experience building products from 0 to 1 - Experience with fine-tuning LLMs, synthetic dataset creation, or benchmark development - Experience working with sensitive or regulated data domains such as tax, accounting, finance, legal, or healthcare - Familiarity with modern AI evaluation and observability tooling How We Define Success in This Role - A strong Senior AI Engineer in this role: - Ships AI features that create real user value - Improves automation, accuracy, and reliability across important workflows - Operates with high autonomy and strong product judgment - Moves quickly without sacrificing reliability or sound engineering practices - Helps the team make better decisions about how to build and evaluate AI systems Stack Context - TaxGPT's core stack: Django, FastAPI, React / Next.js, Go, AWS EKS - Recent expansion into Azure and GCP - Kubernetes via Porter for deployment - PostgreSQL - GitHub for source control Benefits - Medical, dental, vision, 401k + 3% match, life insurance