I'm an AI engineer with 3+ years of experience taking machine learning and generative AI from idea to production: LLM agents, RAG, predictive models, computer vision, and the data pipelines, APIs and cloud infrastructure behind them. I pair a master's in Robotics & AI from ASU with enterprise engineering experience, and I measure my work by impact: $72M in revenue protected, 7,900+ hours automated every month, and tools used by 2,000+ people.
Every system here is running in production, owned by a real team and measured by the number it moves.
Combines call transcripts with account data. I productionized the XGBoost model that now feeds the retention team's outreach queue.
Replaced per-seat Claude and ChatGPT licenses with one platform: token-based usage tracking, privacy guardrails, MCP integrations into internal apps, and Agent Skills packs for each department.
I owned the retrieval pipeline, the evaluation harness and CI/CD for users worldwide.
Handles contract processing that the procurement team used to do by hand.
Scores calls at a volume a human QA team could never sample by hand.
Speeds up how the sales team answers incoming requests for proposal.
Continuously tracks 10 competitors and replaces manual research.
The program behind company-wide adoption of RELAY and everyday AI tools.
VERA, CARL and ROSIE together automate 7,900+ hours of manual work a month.Python · FastAPI · React · GCP
Computer science in India, enterprise engineering at Infosys, a master's in robotics and AI at Arizona State, then AI engineering at Nextiva. Newest first.
GPA 3.83 / 4.00
CGPA 8.49 / 10.00
Computer vision, deep learning, generative models, big-data analysis, constraint solving and robotics, built outside my day job. Code for most of these is on GitHub.
Translates signs to text in real time so people with hearing or speech impairments can communicate more easily. A multi-modal pipeline in Python and C++ built on MediaPipe hand landmarks and OpenCV, running on an ordinary laptop CPU.
A web app that recognizes hand-drawn doodles, built for a Kaggle-style challenge in my master's ML course. A Keras CNN trained on Google's Quick, Draw! dataset classifies sketches as star, sun, moon, rainbow or cloud. Draw live on a canvas or upload an image, and a Flask server returns the prediction.
Detects disease in plant leaves from photos and recommends a pesticide. An OpenCV pipeline cleans each image (resizing, grayscale, noise removal, edge detection, thresholding, contours and segmentation) to isolate infected regions, then a TensorFlow/Keras CNN trained on the PlantVillage dataset classifies the leaf. Wrapped in a Tkinter desktop app.
Real-time detection of lavender balloons in a live webcam feed. A custom-trained YOLOv8 model finds candidates, and HSV colour masking confirms them, which keeps false positives down under changing light.
Classified accident severity and located high-risk zones across 6M+ US accident records using Random Forest and K-Means clustering. Weather, visibility and time of day came out as the strongest drivers of severity.
Processed and analysed customer reviews of gym businesses from the Yelp dataset on a Hadoop and PySpark stack, surfacing trends gym operators can act on and visualizing the results.
A Stable Diffusion web app that creates marketing visuals from text prompts, deployed on Hugging Face Spaces with CI/CD and inference tuned for production use.
Reorganizes rooms, cabinets and items at the lowest cost while respecting ownership, compatibility and capacity rules (5 items per cabinet, 4 cabinets per room, long items only in high cabinets). Modelled in ASP and solved with Clingo, following the ASP Challenge 2019 problem.
Book seats by tier (Normal, Silver, Prime), pay online and get a digital ticket. Admins manage movies, showtimes, seats, sales and feedback over a six-table relational schema.
Connects donors, patients and blood banks: donor registration, live blood stock tracking, request and fulfilment, and an admin dashboard, built to shorten emergency response times.
Tracks books, members and loans.
Robotics and Autonomous Systems (Artificial Intelligence)
Arizona State University · 2025Microsoft Certified: Azure Fundamentals
MicrosoftMicrosoft Certified: Azure Data Fundamentals
MicrosoftConference presentation of plant leaf disease detection research
IEEE conferenceAward for delivery excellence
InfosysComputer Science
Visvesvaraya Technological University · 2020My toolkit plotted like an embedding space: related tools sit near each other. Hover a cluster to focus it.