Computer Science Undergraduate · Full-Stack & Backend Developer
B.Tech CSE @ SVKM's NMIMS (MPSTME), Shirpur · Class of 2028

I build complete, tested, deployed systems: event-driven backends, databases that do real work, and clean frontends on top. My recent work includes a Kafka-based fraud-monitoring platform, a fraud engine written inside Oracle with PL/SQL, a compiler I wrote by hand from lexer to C++ code generation, and a webcam-controlled game built with classical computer vision.

• 🔭 Focus: backend engineering, event-driven architecture, databases, and full-stack web apps
• ✅ How I work: every project is deployed live and backed by automated tests and CI (450+ tests across projects)
• 🎯 Looking for: Software Engineering internship opportunities

FastAPI Apache Kafka PostgreSQL WebSockets React TypeScript Docker Airflow

• Event-driven pipeline: API → PostgreSQL → Kafka → risk engine → WebSocket, pushing APPROVE / REVIEW / BLOCK decisions to the dashboard in real time
• Risk engine scores amount, new-device, transaction-velocity and behavioural-anomaly signals, with explainable reasons
• 6 containerized services behind an Nginx reverse proxy; CI with Pytest, Jest and Cypress end-to-end tests

Oracle 21c PL/SQL ORDS React 19 Playwright

• Fraud detection runs inside the database: a compound trigger scores every transaction, raises alerts and assigns them to the least-loaded analyst
• Normalized 12-table schema with constraints, reporting views and indexes checked with EXPLAIN PLAN
• Authenticated REST API with role-based access, a two-role React dashboard, and 150+ automated checks

Python FastAPI React TypeScript C++17

• Every phase hand-written, no parser generators: lexer, recursive-descent parser, typed AST, semantic analysis, three-address code, optimizer, C++ code generator
• Optimizer verified by a 300-program randomized differential test; 233 tests in total, including g++ build-and-run checks
• Interactive workbench that shows tokens, AST, symbol table, original vs. optimized TAC, and the generated C++ live

Python OpenCV NumPy Pygame

• Steer a spaceship by moving a coloured object in front of your webcam, tracked in ~1.2 ms per frame at 60 FPS
• Classical computer vision only: HSV segmentation, morphology, contour centroids and motion smoothing
• 65 tests using synthetic frames and fake cameras, so no webcam is needed to run them

Languages

Backend & Data

Frontend

DevOps & Testing

• 🥈 2nd Prize, Python Mini Project Competition
• 💡 Smart India Hackathon 2025: Round-1 selection through the Yugantar 2.0 Inter-Institute Hackathon
• 📜 JPMorgan Chase & Co. Software Engineering Virtual Experience (Forage) · Microsoft Learn Cloud Computing badges

Open to Software Engineering internships. If you're hiring or want to talk tech, reach out at ashmita110609@gmail.com.