At Cohere, I worked on North, the platform that lets companies build their own AI agents using Cohere’s LLMs. I helped turn complex retrieval and model workflows into practical tools that teams could deploy, monitor, and scale in real environments. My focus was on making the agent experience more reliable and usable, so enterprises could adopt LLMs without needing deep ML expertise.

At Abridge, I contributed to building Linked Evidence, a clinical AI platform that maps AI-generated summaries back to source audio and transcripts to improve trust and transparency in clinician workflows. I delivered full-stack features using Python, FastAPI, React, and TypeScript, developed secure FHIR/REST integrations with hospital systems, and scaled inference services for real-time and batch clinical documentation. My work supported enterprise EMR workflows and helped reduce documentation friction through reliable, scalable AI-assisted tooling.

At Yext, I helped evolve Yext Answers into an NLP-driven enterprise search platform by building backend microservices in Java and Go to support semantic search, indexing, and high-volume data processing for the Knowledge Graph. I developed front-end features in React and TypeScript for enhanced search experiences and internal tooling, and I improved data ingestion and infrastructure reliability using AWS, Docker, and Kubernetes. My contributions helped scale search relevance and performance for large-scale enterprise deployments.