AI Transformation Engineer

About the company

Eleos Health builds behavioral-health AI that automates clinical documentation, analyzes therapy sessions using speech-to-insights and specialized LLMs, surfaces clinical and compliance insights, and integrates with provider workflows and EHRs.

About this role

<div class="content-intro"><p><strong>Who is Eleos Health?</strong></p> <p>Today, more people than ever are speaking publicly about their mental health. Whether it's ourselves, our friends and family or even public figures, taking care of your behavioral health is no longer a taboo, it's vital, and it's only human. <br><br>Eleos is on a mission to help deliver the world's most effective behavioral care through data, measurement, and personalization. Or simply put, <strong>we want to give clinicians the support they need to do the important work only they can do</strong>.</p></div><h2>What is the Opportunity?</h2> <p>You'll join the AI Platform team, the group building the AI infrastructure, agentic workflows, and tooling that power Eleos's R&amp;D organization. Our mandate is to make the entire software development lifecycle faster and smarter with AI, and we're a small, hands-on team that ships real systems, not pilots</p> <p>This isn't a role where work arrives as tickets, and it isn't a prompt-engineering role. You'll own a set of KPIs the team is driving, and it's on you to turn a fuzzy goal into something measurable: figure out what actually needs to be built, define how you'll know it's working, and build it. That means talking to engineers, understanding where the SDLC breaks down, and turning that into agentic workflows and platform capability that ship and hold up over time.</p> <h2>Who Are You?</h2> <p>You're an engineer who's genuinely built agentic systems, not just used AI coding tools. You're comfortable working across different harnesses (Claude Code and similar are common examples, but you're not tied to one), and you know how to work with agent context: what goes into it, how it's scaffolded, and how to keep it lean and effective as a workflow scales. You understand how agents get designed, evaluated, and made reliable in production.</p> <p>&nbsp;</p> <p>You think and act like an entrepreneur inside the team: given a KPI, you can go find out what's actually slowing engineers down, decide what's worth building, and go build it, without needing the work handed to you pre-scoped. You turn "make it better" into a metric and a way to tell if you actually did. You're comfortable with ambiguity and with owning outcomes, not just tasks. You take ownership without being asked twice.</p> <p>&nbsp;</p> <h2>What You'll Do</h2> <ul> <li>Own a set of KPIs the AI Platform team is driving, and take responsibility for figuring out what moves them, not just executing a backlog.</li> <li>Discover the problem yourself: talk to engineers and teams across R&amp;D, understand where the SDLC is slow or manual, and identify what's actually worth automating.</li> <li>Design, build, and ship agentic workflows that support the R&amp;D SDLC, from prototype through production.</li> <li>Work hands-on across agent harnesses, and design and manage the context, knowledge, and scaffolding those agents run on so they stay reliable as workflows scale.</li> <li>Build the feedback loops that make agents trustworthy: evals, checks, and ways to catch an agent when it's wrong, not just when it's slow.</li> <li>Build and extend the AI platform itself, including agents, integrations, and the infrastructure that runs them.</li> <li>Decide your own work: prioritize what to build next based on impact toward the team's KPIs, not a pre-assigned ticket queue.</li> <li>Prove impact: be able to show, concretely, that what you built moved the KPI you were responsible for.</li> <li>Partner closely with the other engineers on the team to keep the platform coherent as it grows.</li> </ul> <h2>What You Bring</h2> <ul> <li>6-8+ years of hands-on experience in senior software engineering positions.</li> <li>Hands-on experience building and shipping agentic workflows or LLM-powered systems in production, not just prototypes.</li> <li>Real experience working across agent harnesses (e.g., Claude Code and similar), and hands-on understanding of agent context: how it's built, scaffolded, and kept effective at scale.</li> <li>Practical understanding of the actual building blocks: LLMs, RAG, tool use, evaluations, and human-in-the-loop patterns, not just theoretical familiarity.</li> <li>Strong software engineering fundamentals: you can design, build, and operate a system end to end.</li> <li>A demonstrated ability to work without a pre-defined spec: given a goal or a KPI, you can find the problem, scope the solution, and build it.</li> <li>Product judgment: you can separate what's worth building from what isn't, and you know how to tell whether it actually worked.</li> <li>Strong discovery skills: you can sit with engineers, map how work actually happens today, and identify where automation genuinely helps versus where it doesn't.</li> <li>Strong communication skills: you can translate between technical tradeoffs and the people who need to understand the impact.</li> </ul> <h2>Advantages</h2> <ul> <li>A cost-aware mindset: you treat tokens like a resource to be managed, not an afterthought.</li> <li>Experience building on top of Kubernetes, containers, or CI/CD infrastructure in production.</li> <li>Experience with MCP-based tooling, agent orchestration frameworks, or LLM gateways.</li> <li>Familiarity with observability tooling (e.g., Datadog) for monitoring agent or platform health.</li> <li>Background in a fast-paced, high-growth B2B SaaS environment.</li> <li>Exposure to health tech, healthcare institutions, or behavioral health.</li> </ul> <p><br><br></p>