AI Engineer 2, promoted from Software Engineer, AI
Thermo Fisher Scientific · May 2024 – Present
- Shipped a RAG service assistant to production in 2 months as sole engineer, cutting $3.5M in annual costs; used by 3,000+ field engineers; collapsed a multi-step SAP ERP workflow to 1-2 clicks.
- FastAPI, Databricks, PostgreSQL on AWS.
- Scaled an internal document search tool from PoC to 7,000+ daily users, automating ingestion across 400 repositories (SharePoint, S3, Bynder) and raising indexing throughput 7x, from 2 repos/week to 2/day.
- Azure AI Search.
- Designed and built a CPQ sales assistant that turns natural-language requirements into full quotes, routing between a LangGraph multi-agent workflow and a custom deep-research agent; synced to Salesforce CPQ, launching to 600+ reps against $28M in projected revenue.
- Built two revenue-facing agents: a customer intelligence assistant surfacing upsell opportunities across accounts (credited with $3.1M annual revenue), and a lead-scoring pipeline rating scraped tender data 1-5 via an LLM agent on an async worker (generates $2M+ incremental revenue).
- Defined the org's MCP server archetype and safety requirements with the platform team, then shipped two of its earliest production servers, extending two assistants into ChatGPT Enterprise with each UI served as an MCP resource.
- Cut p95 latency 53% (26.5s to 12.4s) by triaging a production P1 and parallelizing document and ticket retrieval.
