About Plenful

Plenful is on a mission to transform healthcare operations from the inside out. Fresh off our $50M Series B and backed by Notable Capital, Bessemer Venture Partners, TQ Ventures, Susa/Kivu Ventures, and other leading investors, we’re building the category-defining AI automation platform that healthcare teams rely on to operate smarter, faster, and more efficiently. Our technology empowers healthcare operators across hospital and health systems, pharmacies and payors to eliminate manual work, reduce administrative burden, and improve compliance, all while unlocking critical revenue to fund programs for their in-need patient populations.

Built by healthcare operators for healthcare operators, Plenful is driven by a deep understanding of the challenges facing today’s care teams. We’re passionate about equipping healthcare workers with world-class tools that deliver real, measurable impact, and we’re proud to serve 100+ leading health systems, pharmacies, and healthcare organizations across the country. If you’re excited to help shape the future of healthcare, we’d love to meet you. Apply now to join our growing team.

About the role

Healthcare's back office still runs on faxes, PDFs and phone calls. AI can take that work off people's plates, but only if it is accurate enough to trust with patient and financial data.

At Plenful, we build agents and models that do that work in production, with people in the loop wherever judgment matters. As a Staff Machine Learning Engineer, you set the technical direction for how we build, train and measure AI across the company, and you stay hands-on while you do it. This is an individual contributor role with no direct reports and broad influence.

What you'll do

You set the direction for how our agents are built

Define the architecture for production agents: tool use, retrieval, state and context management, routing, orchestration, tracing and failure recovery.

Turn what works into patterns other engineers reuse.

Make the calls on build versus buy for models and platform pieces.

You own our post-training strategy

Decide where supervised fine-tuning, preference optimization, distillation and reinforcement learning pay off, and where they do not.

Train and ship LLMs and vision-language models for healthcare tasks, and build the in-house capability to keep doing it.

Prove every gain with evaluation that holds up in production.

You define what good looks like

Set the AI quality strategy: automated graders, calibrated LLM judges, regression suites, human review, failure taxonomies and offline-to-online validation.

Make eval results the basis for release decisions.

You build loops that improve themselves

Design the systems that turn production failures and user feedback into better datasets, evals and models.

Set the standard for dataset versioning, observability and MLOps so every team's work is reproducible and measurable.

You raise the bar for everyone

Co-own the 12-month ML roadmap with engineering and product leaders, balancing near-term wins against longer bets.

Lead technical reviews, mentor Senior engineers, and help us hire.

Build privacy in by design: minimum necessary data, de-identified or synthetic data where it works, and HIPAA-aware practices throughout.

What we're looking for

8+ years in ML or software engineering, including 3+ years leading LLM or agentic systems in production.

Deep applied post-training experience. You have run supervised fine-tuning, preference optimization, distillation or RL end to end and shipped a model that moved a production metric.

You have designed an evaluation strategy for LLM or agent systems and seen it drive real release decisions.

You have architected production agent systems and can name what broke and what you changed. Experience with frameworks like LangGraph, RAG and vector search (for example Pinecone) is expected.

Strong Python and PyTorch, and comfort with distributed training and serving (for example vLLM, FSDP or DeepSpeed).

Real MLOps depth: Docker, Kubernetes, MLflow or equivalent, CI/CD, and experiment and dataset tracking.