A selection of recent engagements across AI, cloud, and optimization. Client names are withheld — the work speaks for the kind of problems I take on.
AI · Data engineering
Designed and built a cloud-native tool that turns raw customer feedback into a dynamic taxonomy tree. Engineered end-to-end pipelines to ingest HubSpot CRM data, then applied embedding-vector clustering and dimensionality reduction to classify it — full-stack Python and React, with infrastructure automated in Terraform.
Optimization · Serverless AWS
Sole technical owner of an end-to-end platform optimizing pickup and delivery schedules. Designed a serverless AWS architecture (Amplify, Lambda, RDS Aurora) and a custom optimization engine on Google OR-Tools running via AWS Batch to handle complex scheduling constraints — React, Flask, Python, Terraform.
AI · Document intelligence
Engineered an AI pipeline that ingests thousands of pages of unstructured insurance data (PDFs, images, OCR) into clean structured records. Built on AWS Step Functions and Textract with a LangGraph orchestration pipeline, validation scoring, and CloudWatch monitoring — delivered with source-highlighting in the PDF UI and an SSR-React frontend backed by Valkey and DynamoDB.
Cloud consultancy
Provided architectural oversight for a large-scale cloud migration: technical risk assessment and mapping to move legacy MS SQL workloads to AWS Babelfish, plus secure Amazon QuickSight dashboards with row-level security for multi-tenant data isolation.