Work Experience
Software Engineer I
AML — Compliance Suite
- Delivered a full compliance suite with 91 APIs — name screening, transaction monitoring alert ingestion, alert lifecycle and case management — in 5 months.
- Built a name screening engine from scratch using RapidFuzz fuzzy matching and weighted multi-attribute scoring in Python and MongoDB Atlas.
- Cut real-time name screening API latency by 87%+ (9s → 1.2s) through algorithm optimization, system architecture, and fine-tuned MongoDB Atlas configuration.
- Redesigned customer rescreening into a parallel SQS packet-worker pipeline and tuned MongoDB using real-time disk throughput, page-fault and connection-pool analysis — cutting per-customer screening time nearly 2,839x (21.6s → 7.6ms) while scaling to 200K+ customers per run.
- Built a centralized knowledge graph across all Python + Go microservices by statically parsing each service to its AST, mapping code entities, relationships and cross-service calls into one queryable graph.
- Cut cross-service impact analysis by 77% cost and 5× latency by exposing the knowledge graph to an LLM via MCP for plain-English queries grounded in real code.
Digital Goods — Marketplace
- Delivered 75+ backend APIs in Go, Postgres and Redis to launch the Digital Goods marketplace, integrating 200+ billers on a platform architected to scale to 500K users and $1M GTV in year one.
- Shipped the backend for a consumer MVP across 12+ user flows in under 3 months, powering a milestone that helped secure $3M+ in funding.
- Built a fault-tolerant, event-driven order-fulfillment flow with AWS SNS/SQS, integrating vendor and payments systems.
- Automated vendor-company reconciliation in Go + AWS Lambda across 10,000+ transactions.
- Strengthened API security with a middleware library blocking XSS/SQL injection, adopted across 5 critical teams (Auth, Payments, KYC, Users, Digital Goods).
DSA Mentor Intern
- Mentored 100+ students in complex data structures and object-oriented design patterns, conducting targeted workshops to enhance engagement and comprehension of advanced programming concepts.
Machine Learning Intern
- Automated real-time IoT data collection for paddy fields by scripting Raspberry Pi devices with Python.
- Improved model training and decision-making by 10% through clustering-based data validation testbeds.
- Achieved 5% accuracy gain and 20% communication overhead reduction by implementing Federated Learning.
Projects
An AI agent that autonomously plans and ships code across 3 microservices via a 5-phase workflow (Discover, Implement, Verify, Ship) with a human-approval gate.
- Built on the Claude Agent SDK with 29 MCP tools, Pydantic contracts and per-run Git worktree sandboxing as guardrails.
- Cut LLM cost 30% and returns cross-service impact analysis in under 90s using GraphRAG over a 7.5K-node Neo4j knowledge graph with model-tier routing and prompt-cache-aware context engineering.
- Improved follow-up answer accuracy with semantic retrieval — conversation turns embedded via the E5 model and ranked by cosine similarity to surface the top-5 relevant exchanges as grounding context.
Attained 70% accuracy in tumor classification by designing a VGG16-based ML pipeline. Deployed a Flask web app with CI/CD using Docker, GitHub Actions, and AWS infrastructure.
Crafted a Spring Boot application with MySQL JWT, improving API response time by 15% through RESTful API. Deployed the microservice on AWS using Docker, achieving 99% uptime via CI/CD pipelines.
Education
Manipal Institute of Technology
B.Tech in Computer Science & Artificial Intelligence
Technical Skills
Programming Languages & Concepts
Frameworks & Libraries
Tools & Technologies
Development Methodologies & Practices
Courses
Coding Profiles
Research Profile
Achievements
Won First Prize at MUJ HackX Hackathon (Oct 2023) for an Al-powered dermatoscope app.
Solved over 750 DSA problems, earning 8 badges and a LeetCode Knight with max rating of 1,965 (Top 3%).
Achieved global ranks of 71 (Weekly Contest 438) and 842 (Biweekly Contest 145) in LeetCode contests.
Secured First Place in a competitive coding event on HackerRank, demonstrating strong algorithmic skills.
Consistently awarded the Dean's List excellence Award for three years (2nd to 4th year) as Branch Topper.