Resume

• Took over as technical owner of the entire platform after the lead developer departed, growing ASK NOHA into an agentic system orchestrated with LangGraph and exposed as governed tools through an MCP server, with human-in-the-loop approval on sensitive actions and a policy layer enforcing autonomy boundaries
• Developed a versioned evaluation harness with bootstrap confidence intervals on retrieval deltas, ablation studies, and an LLM-as-judge calibrated against human labels via Cohen's kappa — used as a release gate, not a report
• Shipped four more production systems end to end: a publishing platform with a role-based permissions model I designed, a multi-channel campaign communications platform used in a real election, and a geospatial field operations dashboard used in a real campaign
• Engineered a canonical MongoDB platform unifying 1.7M+ public-record profiles and ~4M historical participation records, reconciling a 734,457-record snapshot into 20,060 new profiles while isolating a single ambiguous identity for review
• Administer the platform's infrastructure end to end — the public site on Vercel, the summarization service on Azure Container Apps, and everything else, including ASK NOHA's agent runtime and MCP server, on a Linux server I run myself
• Own product design, system architecture, implementation, deployment and operations across the platform, coordinating a team of up to four including a remote developer

• Joined as the platform's AI engineer and built its first AI service: the meeting summarization microservice on Azure Container Apps, plus the transcript acquisition and ASR pipeline with Whisper fallback for meetings without published captions
• Proposed and began building ASK NOHA, a hybrid RAG and knowledge-graph retrieval system over ~1,600 civic meetings, combining BM25, dense retrieval and Neo4j graph traversal with claim-level citation validation

• Turned a founder's early pet-wearable concept into a defined product — shaping requirements, scoping technical feasibility on constrained hardware, and designing the end-to-end system architecture across device, sensing, connectivity and cloud
• Built working prototypes on Raspberry Pi with the Pi AI Camera, running YOLO-based on-device object detection to automatically capture meaningful moments rather than recording continuously
• Developed Python pipelines for real-time ingestion, preprocessing and TensorFlow Lite inference under strict device memory, battery and latency constraints
• Built React dashboards visualizing live telemetry, behavioral signals, device health and model outputs from connected hardware

• Founded and ran a five-line 3D-printing and laser-engraving business spanning workshops, hardware sales, repair services, imported machine distribution and custom manufacturing, managing a team of 4–5
• Built the business's first customer-analytics workflow after inbound volume from IndiaMART reached hundreds of leads per day, moving from manual spreadsheets to Python-based analysis to identify which segments and channels actually converted
• Established the India import channel for Ortur desktop laser engravers, handling supplier relationship, import logistics, pricing and downstream distribution
• Delivered 100+ customer projects across engineering, dental and corporate custom-gift work, translating vague requirements into specifications, production plans, estimates and finished parts

• Taught probability, statistical inference, hypothesis testing and sampling to a cohort of 22 students through structured lessons and applied examples
• Designed lectures, exercises and assessments connecting statistical concepts to real analytical questions
• Explained quantitative methods to students with varied mathematical backgrounds, strengthening communication with non-specialist audiences

Data Science concentration with focus on statistical modeling, machine learning, and applied analytics.