Staff Machine Learning Engineer

Led backend and machine learning delivery for an in-vehicle conversational assistant. Built Python and FastAPI services that convert voice intent into safe, traceable actions across vehicle controls and diagnostics, with agent and tool orchestration through gRPC, REST, and GraphQL APIs.

Integrated Gemini through Google Vertex AI for reasoning while maintaining audited service-side execution controls. Built RAG capabilities with LangChain and Pinecone, created Kafka-to-S3 and PostgreSQL telemetry pipelines, trained PyTorch models with DeepSpeed and FSDP, optimized inference for real-time latency, and deployed observable, secure services on EKS using Docker, Terraform, GitHub Actions, OpenTelemetry, CloudWatch, Prometheus, and Grafana. Mentored engineers and partnered with product, SRE, and platform teams.

Built DLRM-style advertising ranking models in PyTorch using sparse embedding tables and dense features to predict CTR and CVR and improve ad ordering. Used distributed PyTorch training, mixed precision, checkpointing, reproducible configurations, and clean data sharding.

Improved ranking quality through feature and embedding optimization, offline evaluation, and online A/B testing. Developed data-quality guardrails, optimized serving latency through batching and caching, connected regression gates to releases, monitored model quality in production, and mentored engineers on ranking, embeddings, model debugging, and launch discipline.

Built the Expertise Engine real-time agent-assist product using Python, FastAPI, and Node.js to deliver live coaching and next-best-action suggestions for voice and chat interactions. Trained and shipped PyTorch NLP models using transformer architectures including BERT, GPT-2, and T5 for intent, objection, and moment detection.

Developed real-time gRPC inference and routing flows, custom ASR and entity-recognition modules, and secure REST and GraphQL orchestration APIs. Built React, React Native, and Flutter product surfaces; operated Kafka and S3 data pipelines; and deployed Dockerized Kubernetes services using Terraform and GitHub Actions with comprehensive production monitoring.

Designed and implemented high-accuracy NLP classifiers using state-of-the-art deep learning techniques. Led strategies for obtaining high-quality training datasets to support machine learning development.

Defined and analyzed product metrics and built dashboards for company health and enterprise campaign performance. Mentored team members on machine learning projects and engineering practices.

Worked in a fast-paced agile team to iterate on proof-of-concept products for potential partners. Built complex search and recommendation systems spanning API services, backend data storage, search-query logic, and frontend web applications.

Conducted controlled experiments and A/B tests to measure engagement lift relative to competing products. Contributed across the stack to turn data-driven capabilities into partner-facing product experiences.

Built user feature-modeling and engagement-prediction capabilities using machine learning. Led the analytics backend for partner-facing dashboard products.

Developed analytics systems for high-throughput, real-time data streams and mined large datasets for business insights and data-driven decisions. Applied machine learning to product re-engagement and customer upsell initiatives.

Worked as a full-stack developer on a web application designed to significantly improve the trip-planning experience.

Also developed an iOS version of the application, extending the product experience to mobile users.

Developed software infrastructure to optimize simulation performance for CPU core verification.

Contributed engineering solutions supporting processor verification workflows and simulation efficiency.

Master of Science in Artificial Intelligence.

Bachelor of Science in Computer Science.