Job description

About Us Savant is transforming how healthcare and life sciences organizations unlock the value trapped in unstructured medical data. Our platform combines cutting-edge large language models (LLMs) with domain-specific quality controls to convert free-text clinical records into structured, analysis-ready data — efficiently, accurately, and at scale. We work with leading institutions across healthcare, life sciences, and research, supporting faster studies, sharper insights, and better care. Backed by Roivant (NASDAQ: ROIV), Savant is built for organizations that see structured data not just as an output, but as a foundation for innovation. The Opportunity We’re hiring an AI Engineer to improve the quality and efficiency of Savant’s clinical data platform. This is a generalist applied-AI role for someone who can move fluidly between software engineering, data science, experimentation, and production operations. You’ll study how changes throughout our pipeline affect system performance, then design and ship improvements. That may mean changing how we prompt or structure interactions with LLMs; how we retrieve, rank, and assemble supporting information; how workflows are decomposed and controlled; or how outputs are validated and repaired. If you’re excited by the challenge of making AI systems measurably better, enjoy following evidence across the entire stack, and want your experiments to become real product capabilities, we’d love to hear from you. Savant is based in NYC, and this role is remote-eligible. Preference will be given to candidates able to come to work in NYC. Some Things You Might Work On Establish software programming paradigms for our platform’s interactions with LLMs in both token-parsimonious and token-hungry scenarios Build a test bench for understanding the $/performance frontier of open- and closed-weight language models as relates to our platform’s (often very different) core workloads Partner with clinical experts to encode domain knowledge and define meaningful quality standards Build evaluation suites that measure accuracy, completeness, consistency, latency, and cost Design controlled experiments that isolate the effects of changes throughout the platform Improve model interactions through prompting, structured outputs, tool use, context management, and workflow composition Build datasets and internal tools for evaluation, regression testing, and rapid iteration What We’re Looking For 5+ years of experience in applied machine learning, AI engineering, data science, or adjacent software engineering roles Excellent Python and SQL skills Hands-on experience building production applications with LLMs Experience designing experiments and interpreting noisy, multidimensional results Familiarity with LLM evaluation, retrieval-augmented generation, agentic workflows, or MLOps Strong software-engineering fundamentals and the ability to ship reliable production code Intellectual curiosity, sound judgment, and taste Bonus Points For Knowledge of statistical inference, experimental design, or human-in-the-loop evaluation Experience building evaluation platforms, experiment tooling, or model-observability systems Familiarity with information retrieval, search, ranking, entity resolution, or knowledge systems Experience with clinical data, medical terminology, or other healthcare data Familiarity with HIPAA, de-identification, or quality controls for high-stakes data Experience in an early-stage startup or small, fast-moving technical team Why Join Us Systems-Level AI Work: Improve the entire applied-AI system, not just individual prompts or models Rigorous Iteration: Build the evidence and evaluation infrastructure needed to distinguish real progress from anecdotal wins Production Impact: See experiments become capabilities used on consequential healthcare data Broad Ownership: Work across models, retrieval, data, evaluation, infrastructure, and product behavior Mission-Driven Work: Help make messy clinical information as useful and dependable as structured data Base salary for this role will be determined during the interview process and will vary based on multiple factors, including but not limited to prior experience, relevant expertise, current business needs, and market conditions. The expected base salary for the role will generally be between $175,000 to $250,000 per year, with meaningful equity. The final salary offered may be outside this range based on individual circumstances and business and market conditions. Base salary if hired is only part of the total compensation package, which, depending on the position, may also include other components such as discretionary bonuses and Company-sponsored benefit programs. This position is at-will and Savant reserves the right to modify base salary (as well as any other discretionary payment or compensation program) at any time, including for reasons related to individual performance and business and market conditions. Savant provides equal employment opportunities to all employees and applicants for employment and prohibits discrimination and harassment of any type without regard to race, color, religion, age, sex, national origin, disability status, genetics, protected veteran status, sexual orientation, gender identity or expression, or any other characteristic protected by federal, state or local laws.

Requirements

Must-haves

We’re hiring an AI Engineer to improve the quality and efficiency of Savant’s clinical data platform.
This is a generalist applied-AI role for someone who can move fluidly between software engineering, data science, experimentation, and production operations.
If you’re excited by the challenge of making AI systems measurably better, enjoy following evidence across the entire stack, and want your experiments to become real product capabilities, we’d love to hear from you.
Savant is based in NYC, and this role is remote-eligible. Preference will be given to candidates able to come to work in NYC.
Some Things You Might Work On
Establish software programming paradigms for our platform’s interactions with LLMs in both token-parsimonious and token-hungry scenarios
Build a test bench for understanding the $/performance frontier of open- and closed-weight language models as relates to our platform’s (often very different) core workloads
Partner with clinical experts to encode domain knowledge and define meaningful quality standards
Build evaluation suites that measure accuracy, completeness, consistency, latency, and cost
Design controlled experiments that isolate the effects of changes throughout the platform
Improve model interactions through prompting, structured outputs, tool use, context management, and workflow composition
Build datasets and internal tools for evaluation, regression testing, and rapid iteration
What We’re Looking For
5+ years of experience in applied machine learning, AI engineering, data science, or adjacent software engineering roles
Excellent Python and SQL skills
Hands-on experience building production applications with LLMs
Experience designing experiments and interpreting noisy, multidimensional results
Familiarity with LLM evaluation, retrieval-augmented generation, agentic workflows, or MLOps
Strong software-engineering fundamentals and the ability to ship reliable production code
Intellectual curiosity, sound judgment, and taste
Knowledge of statistical inference, experimental design, or human-in-the-loop evaluation
Experience building evaluation platforms, experiment tooling, or model-observability systems
Familiarity with information retrieval, search, ranking, entity resolution, or knowledge systems

Nice-to-haves

Bonus Points For

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aimachine learningpythonsqlllm

Platform: greenhouse
Effort: quick