AI Engineering · Portfolio

Revin AI Engineer

I build the parts of model systems that decide whether they survive contact with production — retrieval, evaluation, and the unglamorous plumbing between a demo and something a team can rely on.

Focus
AI systems — retrieval, evaluation, inference
Stack
Python · PyTorch · Postgres · FastAPI
Open to
Full-time · Contract
Status
Open to work

Selected work

05 projects
01 Retrieval over a legal archive Hybrid search · Reranking · Evals

Tuned a retrieval layer across 400k documents, then treated quality as an engineering problem: chunking strategy, reranking, and explicit refusal behaviour all tracked as a regression suite.

02 An evaluation harness for agent runs Evals · Tracing · CI

Made agent behaviour measurable: recorded real runs, scored tool selection and task completion, and gated merges on the score so regressions surfaced before users found them.

03 Small models on edge hardware Quantization · ONNX · Latency

Compressed and served a small model for offline use, trading accuracy against latency and memory until it ran acceptably on hardware with no GPU and no reliable connection.

04 A fine-tuning loop that stays honest Fine-tuning · Data curation

Built the boring half of fine-tuning: dataset construction, leak checks, and a held-out set that changed how we judged every subsequent checkpoint.

05 Prompt regression testing as a habit Tooling · Pytest · Diffs

Turned prompt edits into reviewable diffs with pass/fail cases, so changing a system prompt stopped being an act of faith.

Capabilities

04 areas

Most model work fails for ordinary reasons — bad retrieval, no measurement, costs nobody watched. That is the part I take seriously.

Building the actual product around a model: tool use, structured output, retries, fallbacks, and guardrails that fail loudly instead of quietly.

Ingestion, chunking, embeddings, hybrid search, reranking — and knowing which of those actually moved the number before adding the next one.

Datasets that reflect real traffic, scoring that a teammate can argue with, traces that explain a bad answer instead of just recording it.

Serving, caching, batching, quantization and routing. Latency and spend treated as product decisions, not as surprises on an invoice.

Notes

Writing
2026 · 02 Refusal behaviour is a product decision, not a safety checkbox ↗ 2025 · 11 What I check before adding a reranker ↗ 2025 · 08 Fine-tuning is a data problem wearing a model costume ↗
Open to roles & collaborations
hello@revin.dev Copied