Technical Skills

Programming Languages

Python, Javascript, Java, C++, Rust, SQL

Backend Frameworks

FastAPI, Async I/O, Postgres, Redis, pytest, boto3, Django, Spring Boot, JPA, MongoDB

Frontend Frameworks

HTML, CSS, React, WebAssembly, React Testing Library

Cloud & DevOps

AWS, Azure, GCP, Docker, Kubernetes, Terraform, GitHub Actions CI/CD, Git

Certifications

AWS Cloud Practitioner, Azure Data Fundamentals

Education

GRADUATED MAY 2025

Master's in Computer Science

Virginia Tech

GRADUATED MAY 2020

Bachelor's in Computer Engineering

Pune University

Experience

AUG 2024 – DEC 2024

Testbed Intern

Commonwealth Cyber Initiative

  • Researched and implemented Python FastAPI interface for O-RAN O-Cloud standards. Wrote documentation and communicated technical updates to team and telecom industry stakeholders.
  • Deployed to a Kubernetes cluster using Docker images and YAML manifests, managing rollouts and basic scaling for reliable application delivery.

OCT 2021 – JUL 2023

Digital Specialist Engineer

Infosys Limited (Client: Bloomberg Media)

  • Developed, debugged and maintained full-stack Python FastAPI/React RESTful user-facing APIs with PostgreSQL, deployed on AWS (EC2/S3), maintaining 99.9% uptime.Redesigned SQL schemas to implement indexing, cutting API latency by 20% for high-traffic endpoints.
  • Optimized legacy AWS-based AI training job for cost-efficiency and scalability, redesigning it to to a serverless event-driven AWS architecture using Lambda, S3, Sagemaker, and ECR; reducing infra cost by 40% .
  • Setup CI/CD using GitHub Actions to automate deployment for frontend and backend components and containerized them using Docker, making deployment 50% faster.
  • Set up application monitoring and log collection using AWS CloudWatch metrics, alarms, and centralized CloudWatch Logs, enabling early detection of performance issues and reducing troubleshooting time.
  • Contributed to React Web Application using modern JS patterns to create reusable UI components (forms, navigation menus); debugged UI issues using profiling tools.
  • Integrated RESTful web service calls in UI for POST, GET and DELETE and deprecated unneeded API calls.<
  • Integrated ONNX Runtime + TensorRT for transformer-based model inference, reducing GPU memory fragmentation and cutting response latency by 2× on NVIDIA T4 instances.
  • Built a RAG pipeline leveraging early LLaMA-7b and zero-shot prompting; tuned HNSW indexing for low-latency document retrieval across 100K+ corpus entries in pre-production.
  • Iteratively developed Python PoCs for ambiguous business use-cases by applying advances in AI/ML, gathering feedback from VP and client stakeholders; presenting live demos and securing budget for full-scale roll-outs.

JAN 2021 – OCT 2021

Programmer Analyst Trainee

Cognizant (Client: Merck)

  • Contributed to code & design review in the fast-paced agile sprints delivering 10+ features for Django/React app with MySQL to automate the validation and ingestion of manufacturing batch data.
  • Automated SQL queries using Python and SQLAlchemy, eliminating manual effort and report generation and accelerating the delivery of key business reports to stakeholders from three days to a single day.
  • Diagnosed and resolved complex issues by analyzing the frontend (UI, API integrations, analytics), reducing issue resolution time by 30% through root cause analysis and collaboration with teams.
  • Optimized rendering of large React tables (>5K rows) using memo and debounced filters. Analyzed performance via Chrome and Firefox DevTools, cutting time-to-interactive from 4.2s to 1.1s and reducing memory usage by 40%.

Projects

Graduate Capstone: L.E.A.R.N

An AI-assisted learning platform that generates personalized K-12 lesson plans using the OpenAI API and RAG. Spring Boot backend handled role-based auth, AI orchestration, and file storage via AWS S3.

Python AI AWS RAG React Spring Boot

Political Bias Detection

Finetuned Llama 3 8B and Gemma 7B with QLoRA on a POLITICS subset to detect and explain political bias; best Gemma 7B q5_k_m model achieved ~75% accuracy.

Python PyTorch LLMs llama.cpp

Chip-8 Emulator

A virtual CPU built in Rust, decoding 35+ opcodes with timers, stack, memory, and graphics rendering; includes a WebAssembly port to run games in-browser.

Rust WebAssembly

System Stats Monitor

A modern, memory-safe system monitor in C++ delivering low-latency real-time updates with under 15 MB RAM usage for resource-constrained systems.

C++

BE MCQs Site

Temporary exam-material crowdsourcing site built with Flask and HTML/CSS; served ~15,000 users and ~250k requests in a little over a month on an Azure VM behind Nginx and Cloudflare CDN.

Python Flask Azure VM Cloudflare Nginx Git

Soccer Data Scraper and Visualizer

Scraped and transformed Wikipedia data for 4 football leagues across 100+ years and produced 35+ season visualizations for analysis using Plotly.

Python Plotly pandas BeautifulSoup Git