LLM Autopilot: Unattended AI Workflow

Category :

GenAI & LLMs

Dates :

Jan 2025 - Present

A scheduled LLM pipeline that runs without supervision and fails safely.

Validation gates, retries with backoff, file locks, and catch-up logic keep failed or malformed runs from reaching users.

Integrated Python services with external APIs, Docker, and Linux scheduling, using logs and validation stages to diagnose failed runs.

More projects

GenAI & LLMs

A retrieval system over 10,000+ documents, tuned and measured like a real product.

GenAI & LLMs

Full-stack agentic AI platform for job seekers, built as sole engineer.

GenAI & LLMs

Scores and ranks candidates against a job using structured extraction, semantic retrieval, and LLM scoring.

Robotics

One interface for running experiments on simulated and physical Unitree Go1 robots.

GenAI & LLMs

Turns live meeting audio into speaker-labeled transcripts, summaries, decisions, and action items.

GenAI & LLMs

Grounded, auditable question-answering over financial and regulatory documents.