AI and machine learning

The full machine learning lifecycle. Training, deployment, and the systems that keep models honest in production.

01 / Training and fine-tuning

Supervised and RLHF fine-tuning of open-weight models. Custom datasets, curriculum design, and eval harnesses from day one.

02 / MLOps and deployment

Model serving with quantization, batching, and autoscaling. Versioning, rollback, and the unglamorous work that keeps inference up.

03 / Applied AI and LLM products

Retrieval pipelines, agent systems, function calling, and structured output. We build the production stack, not the demo.

04 / Computer vision and multimodal

Image classification, object detection, OCR, and multimodal models. From dataset curation through training to inference at the edge or in the cloud.

05 / Evals and governance

Offline evals, online A/B, drift monitoring, red-teaming. Models in production need the same engineering discipline as any other system.