Vyoma Gajjar

I build AI systems across predictive machine learning, generative models, and tool-using agents.

My work connects model behavior to real outcomes across data, modeling, evaluation environments, deployment, monitoring, and production feedback loops. I publish the code, evidence, and negative results behind current experiments.

featured research engineering

agent-economics-lab

Can an agent decision survive missing evidence, weak graders, and the cost of the work it creates?

This open-source, standard-library-only Python project audits whether the evidence around an agent supports a decision to expand it, supervise it, or switch it off. It combines trace-to-outcome evaluation, full-cost accounting, label-quality audits, and reproducible claims. It also publishes corrections when a result does not survive its own checks.

10
public agent datasets audited under one protocol
31,389
runs carrying both outcome and proxy signals
588
controlled required-gate perturbations

These are dataset audits, controlled software experiments, and one small paired public case. They are not model rankings, production prevalence estimates, or deployment recommendations.

reproducible artifacts

questions I am working on

  • How should predictive models, foundation models, and tool-using agents share a production feedback loop without hiding distinct failure modes?
  • Which changes improve end-to-end behavior: data, model, retrieval, tool interfaces, routing, or control policy?
  • When outcomes and cheap proxies disagree, which claims survive, and where should the system abstain or hand off?

work across the stack

machine learning
NLP, predictive modeling, ranking, monitoring, drift detection, and governance.
generative AI
Retrieval, synthesis, evaluation, guardrails, and observability for systems built on foundation models.
agent systems
Tool use, planning, routing, recovery, human handoffs, and the conditions under which a system should earn more autonomy.

technical notes and writing

external receipts

A few things the internet can verify without trusting this page.

Machine-readable evidence ledger with exact claims, immutable project sources, a reproduction command, source classes, evidence locators, and verification dates.

speaking

I occasionally speak about production machine learning, generative AI, agentic systems, evaluation, and responsible deployment.

For talks: gajjar.vyoma@gmail.com

teaching

I teach practical AI systems topics through public courses and workshops.

about

I am a software-engineering-first applied AI engineer. My technical work spans NLP and predictive models, RAG pipelines, and tool-using agents.

The through line is production: how to make models measurable, observable, steerable, and useful once they leave the notebook.

I am interested in systems that learn from feedback without hiding failure, and in teams that value empirical questions, reproducible artifacts, and clear negative results.

San Francisco Bay Area.

Email LinkedIn GitHub