Technologies: AI agents, LLM inference (self-hosted / open-weight), Python, TypeScript, GPU infrastructure, voice AI, data pipelines
Résumé/CV: https://ferretout.ai (live product — status page at /status is the résumé)
Email: sfiejgit2020@gmail.com
Seeking contract / fixed-scope work, not full-time. Specialty: AI agent systems that keep working when nobody's watching — the agent logic, the GPU infrastructure underneath, and the guardrails.
Live proof instead of a portfolio claim: my product FerretOut ingests the full Hacker News firehose 24/7 and scores it for buying intent on optimized open source models — the status page at https://ferretout.ai/status is public and running right now.
What I take on, fixed scope: agent/automation builds that run unattended · open-weight LLM serving (quantization measured against honest evals, batching, failover; private inference for data-residency/HIPAA work) · voice agents for children and elderly users (sub-second turns, guardrails) · high-volume data pipelines (ingest, enrich, classify).
How I work: written scope doc before anything is billed, 50% to schedule, async progress you can check at any hour — no meetings needed.
Remote: Yes (async-first, worldwide)
Willing to relocate: No
Technologies: AI agents, LLM inference (self-hosted / open-weight), Python, TypeScript, GPU infrastructure, voice AI, data pipelines
Résumé/CV: https://ferretout.ai (live product — status page at /status is the résumé)
Email: sfiejgit2020@gmail.com
Seeking contract / fixed-scope work, not full-time. Specialty: AI agent systems that keep working when nobody's watching — the agent logic, the GPU infrastructure underneath, and the guardrails.
Live proof instead of a portfolio claim: my product FerretOut ingests the full Hacker News firehose 24/7 and scores it for buying intent on optimized open source models — the status page at https://ferretout.ai/status is public and running right now.
What I take on, fixed scope: agent/automation builds that run unattended · open-weight LLM serving (quantization measured against honest evals, batching, failover; private inference for data-residency/HIPAA work) · voice agents for children and elderly users (sub-second turns, guardrails) · high-volume data pipelines (ingest, enrich, classify).
How I work: written scope doc before anything is billed, 50% to schedule, async progress you can check at any hour — no meetings needed.
Portfolio: https://contra.com/jackie_yu_tks71b45