I help teams take AI and machine learning from prototype to reliable, deployed systems — with an emphasis on correctness, auditability, and deployment in demanding environments. Deep roots in healthcare and life sciences, and a growing book of enterprise AI work.
A decade-plus of hands-on engagements across the industry — delivered end to end, from research through production.
Full-arc capability across a project — from modeling strategy to systems that run in production.
Auditable, human-in-the-loop AI systems that pair deterministic computation with LLM judgment — and deploy on-prem or fully offline in regulated clinical and enterprise settings.
Genomic and variant interpretation, genotype-to-phenotype and drug-resistance prediction, and protein language models for sequence-based property and specificity prediction.
Rigorous modeling with the small, imbalanced, and expensive-to-label data that real work produces — from biomedical to enterprise — and honest assessment of what the data can and can't support.
Production software and data engineering, cloud platforms, APIs, and agent orchestration — the pipelines and plumbing that make AI and data systems dependable enough to ship.
A few principles that hold across every engagement, especially the high-stakes ones.
The arithmetic and rules are computed, not guessed. AI is reserved for judgment, so results stay correct and reproducible.
Every recommendation traces back to its evidence. The reasoning path is the explanation — no black boxes in high-stakes decisions.
Systems designed to run on-prem or fully offline, inside the security and compliance constraints of regulated environments.
PhD-led delivery that carries the work from strategy through a live, tested, deployed system — with clear accountability throughout.
Where AI and ML can actually help, what's feasible, and the right architecture and roadmap — including data readiness and model feasibility — before committing to a build.
End-to-end design and delivery of ML systems, data pipelines, and cloud infrastructure — from proof of concept to deployment.
Embedded technical guidance on ML strategy and system design, plus model and architecture review and the scientific and grant writing behind funded research.
TRM Systems is the consulting practice of Theodore Mellors, a machine learning engineer with a PhD from Dartmouth's Thayer School of Engineering.
Over the past decade, Ted has built machine learning, data, and AI systems across biotechnology, pharmaceuticals, and healthcare, helping organizations translate research into production software. His experience includes leading machine learning for a Medicare-covered diagnostic, developing AI and data systems for a global biopharmaceutical company, and building production infrastructure for large healthcare organizations.
Today, TRM Systems focuses on agentic AI, machine learning, data engineering, and scientific software. Recent work includes auditable AI agents, genotype-to-phenotype and drug-resistance modeling, protein and foundation-model applications, production data platforms, and technical writing supporting funded research.
The goal is straightforward: build reliable, maintainable AI systems that solve real problems and are ready for production.
Tell me what you're working on. I'll tell you honestly whether and how I can help.
Start a conversation