Software Engineer II at Microsoft on Azure Resource Graph, building the distributed systems behind 12B+ unique Azure resources every day — and scaling to 20B+. Before that, I shipped production multi-agent AI at Nissan and high-throughput ad-tech data pipelines at Amazon. I build cloud-scale backends, distributed systems, and data pipelines that stay fast, reliable, and correct at massive scale.
Also, while you are here, do you want to make your job search faster? Use the Job Search Tool I created to query and filter opportunities quicker and more efficiently.
Experience
At Microsoft, I work on Azure Resource Graph, saving 12B+ unique Azure resources every day and aiming to scale to 20B+ over the next year.
Tech Stack: Azure, GitHub Copilot
At Nissan, I designed and shipped Tell Me More, a production multi-agent agentic chatbot integrated into Nissan's dealership appointment pipeline. Architected a four-agent system on AWS Bedrock AgentCore with persistent session memory, handling multi-concern diagnostic intake flows end-to-end. Owned the full AWS infrastructure, chat application backend, and partner integrations — successfully onboarding 2 dealership channels.
Tech Stack: AWS (Bedrock AgentCore, ECS, Step Functions, DynamoDB, API Gateway, CloudWatch, S3), Multi-Agent Systems, Agentic AI, Python, FastAPI, TypeScript, Serverless Architecture, CI/CD
At Amazon Advertising, I developed scalable backend services and high-throughput data pipelines using AWS technologies including ECS, Lambda, DynamoDB, Step Functions, Glue, and Athena. I built robust ad labeling systems that support both pre- and post-bid filtering to improve ad relevancy, while separately enhancing brand safety systems to protect advertisers and ensure compliance. My work enabled efficient large-scale processing and evaluation of ad data, significantly boosting overall ad quality.
I also led seamless service transitions and implemented automation to improve operational reliability and customer onboarding.
Tech Stack: AWS (Lambda, EC2, Fargate, Batch Jobs, RDS, SNS, SQS, Step Functions, DynamoDB, Bedrock, Kendra, CloudWatch, Glue, Athena, ElastiCache, S3), Distributed Systems, Apache Iceberg, GraphQL, Kafka, Serverless Architecture, Machine Learning, CI/CD, Observability, Java, Python, TypeScript
At Openprise, I developed API integrations with Google Calendar and built ETL pipelines with data preprocessing for PostgreSQL database ingestion and optimization.
Tech Stack: Python, REST APIs, PostgreSQL, ETL Pipelines
At UMBC, I mentored students in Information Retrieval, Computer Security, Discrete Structures, and Numerical Computations courses. Researched fake news detection using machine learning and NLP at AI-SSGL lab. Developed automation scripts for grading and maintained research lab website.
Tech Stack: Python, Machine Learning, NLP, WordPress, Automation
Revamped ML test suites and built preprocessing tooling to accelerate NLP and unsupervised workflows, reducing runtime and manual effort. Created PyQt utilities for automated data preparation and visualization and prototyped ML solutions to improve software testing and evaluation.
Tech Stack: Python, Keras, PyQt, Machine Learning, NLP, Data Visualization, Test Automation
Education
Key Projects
A production four-agent agentic system that triggers when a Dealership customer books a service appointment — the dealership sends a personalized link and the system conducts a structured diagnostic intake conversation before the visit. The Conversation Agent manages customer dialogue and a priority-ordered concern queue and orchestration with the other agents as tools; the Rubric Agent performs real-time structured field extraction and answer quality assessment; the Vehicle History Agent pulls past service records to surface relevant repair context; and the Dealership Internal Agent consults internal systems for warranty coverage, parts availability, and dealer-specific advisories. Each concern gets its own full intake cycle with persistent agent memory across turns. Coordinated integrations with internal X company services and five dealership partners.
Built a full-stack diabetes prediction app using a Random Forest model with Google Maps and Twilio for real-time alerts. Dockerized and deployed on AWS EC2 with PostgreSQL.
Developed an Android app (Java) for a football fantasy tournament and a Flask REST API + scraper to provide live data and scoring.
Architected a P2P distributed system with replication and version control to enable resilient file CRUD operations across nodes.
Technical Skills
Distributed systems design, high-throughput data pipelines, serverless architecture, API development and modeling, ML/AI integration, infrastructure and performance optimization, real-time event-driven systems, observability, multi-agent system design, and prompt engineering.