At Datasite, I implemented observability, evaluation, and quality control for production GenAI and ML models, including LangSmith tracing for OpenAI model calls and pipelines for human audits.
At Bluecrew, I led the MLOps team’s work on deploying models for scalable real-time inference in Databricks. I also worked with a team on an LLM chat interface to collect work history and create professional resumes.
At Conversica, I led a redesign of a question-answering feature built on large language models, reducing the time needed to add questions by 90%. I also developed NLP and ML models that automated more than 100,000 email responses per week and built CI/CD for text classifiers.

