Diam Hamstra

Work on multiple projects using Cursor, Google Gemini, and Edu Arena to generate and evaluate complete codebases, reviewing the quality of LLM output. Evaluate and review LLM-generated data trajectories and project outputs, assessing reasoning quality, tool-use accuracy, and adherence to task instructions.

Processed and structured large volumes of raw data from complex documents for use in ML training and evaluation pipelines. Designed structured data extraction workflows in Python, defining output fields, data types, validation rules, and handling logic.

Built Python backend services and data pipeline components using FastAPI, processing and transforming training data for ML model evaluation workflows. Developed and maintained Python scripts for data validation, transformation, and quality assurance across ML pipeline stages.

Worked on multiple projects using Cursor and Google Gemini and Edu Arena to generate and evaluate complete codebases, leading a team responsible for reviewing the quality of LLM output. Used Node.js and Python to build small apps for each project, test the code Gemini produced, and automate parts of the review process.

Contributed to frontend development using React.js and Next.js with Storybook, building consistent, accessible UI components used across the mortgage application workflow. Integrated Microsoft Application Insights for performance monitoring and wrote Jest test suites.

Developed and trained NLP models using PyTorch and spaCy for intent classification, entity extraction, and context-aware response generation. Deployed models as AWS SageMaker endpoints and optimized inference latency for production serving.

Built the frontend for Cisco's CLX-LAB virtual machine management system using React.js and Apollo Client (GraphQL). Designed and integrated Lua scripting support into the CLX-LAB-API layer, enabling automation and adaptability for power users.

Built a real-time data analytics platform processing large volumes of streaming data using Kafka and AWS Kinesis for ingestion and Spark for distributed processing. Managed deployment with Docker and Kubernetes and used AWS Glue for ETL pipeline orchestration.

Developed an AI-driven mobile application in React Native to detect crop diseases from photos and connect farmers with local pesticide suppliers. Built serverless backend with Node.js, AWS Lambda, DynamoDB, and API Gateway.

Developed and maintained the AOM private frontend framework, a Bootstrap/Foundation-inspired component library used internally across multiple products. Built AOM's custom deployment tool in Node.js to automate build and release processes for 100+ interdependent components.

Bachelor of Technology in Computer Science from Maulana Abul Kalam Azad University of Technology, Kolkata, completed in June 2018.