Hey, I'm Harsh
I am passionate about building scalable applications. I am interested in SDE and Quant roles where I can work on challenging projects and learn new technologies.

Skilled programmer with a knack for creating complex server side applications.
2+ years deep in Competitive Programming
B.Tech in Computer Engineering from DJ Sanghvi College of Engineering, Mumbai (CGPA: 9.21)
CFA Level 2 Candidate
Experience
Here are some work experiences where challenges were turned into accomplishments, making things happen.
Software Engineer
Developed a 2-fold extraction layer to scrape startup data from multiple data sources. Integrated conversational intelligence with Twilio API for adaptive choice-based interviews with founders. Constructed a transformation layer to analyze data chunks and generate hidden insights by leveraging LLMs. Deployed 2 production-grade internal tools for the HR and Marketing team.
SWE Intern
Upgraded an augmented asynchronous client API using in-memory cache reducing the overhead by 22%. Integrated 4 comprehensive and interactive highcharts that display real-time data of agent exposure. Migrated the User-Interface (2 complex screens) with upgraded angular to a mono-repo with over 40 projects. Developed unit angular tests for multiple components of 1st screen thereby extending the test coverage to 82%.
DSA Instructor Intern
I was responsible to coach DSA topics like 2 pointers, graphs, DSU to over 150+ lower silver and bronze level students. I also delivered 5+ high quality problems for biweekly and annual contests.
Sunpy (Open Source)
The community-developed, free and open-source solar data analysis environment for Python. Refactored the sunpy attributes and the world coordinate system to updated version in PR-6732. Overhauled the HTTP error handling in the scraper client along with re-query limiter and extending the pytest coverage to 72% in PR-7541.
Encache (NPM Package - 600+ downloads)
A lightweight, In-memory Async Caching Library for your Nodejs Server. Developed a GET method with 0.01 ms average latency, optimized for high-frequency requests. Integrated 6 eviction policies with seamless dynamic swapping, ensuring no data loss. Implemented LZ4 compression, reducing data chunk size by up to 50%.
Imgress
A stop solution to create, manage and deploy and automate the process of creating your image search engine instances powered by highly accurate vector search databases and img2-vec neural networks layered above resnet50. It comes with highly efficient microservices architecture with serverless computing
Remote Code Execution (RCE) Engine
A remote code execution engine built with NodeJs , supported by docker which is a server-worker-microservice architecture based code execution engine. It performs code virtualization , parellel code execution along with code isolation by code containerization.
SoulSupport.ai (Won UST D3code)
AI based virtual behavioral companion to support a mentally stressed and disturbed individual with text and audio support 24/7. Designed an AI bot utilizing a fine-tuned Falcon-7B model with QLora, enabled with chained conversational responses. Incorporated 6 business features like doctor search, automated user report generation and personalized diet plan recommendations. Built real-time stress and anxiety detection using dual sentiment analysis models processing audio and video data.
Codeforces
Expert, India top 95% percentile in competitive programming.
UST Global D3code 2023
Winners of India's largest hackathon, with over 12500+ participating teams.
TIAA T3Hacks 2024
2nd Runner up after competing against 2200+ registered participants
ICPC Asia West Regionals
3x regional finalist with all India ranks 61, 57 and 84 from 2022 - 2024
Newton School CodeRush March 2023
Global rank 76 out of 12000+ participants
VOIS International Finalists
India top 10 out of 5000+ competing teams
Blogs & Articles
Trying to give back to the community for the greater cause .
Contact Me
Feel free to reach out for collaborations , opportunities or just a friendly hello!
Email: shahh8138@gmail.com