Vaishnavi Bisen — Portfolio

About Me

Education & Background

A passionate engineer who loves turning complex problems into elegant, scalable solutions.

I'm a final-year B.Tech student in Electrical Engineering at IIT Delhi with a deep interest in software development, machine learning, and systems programming. I enjoy working across the full stack — from low-level C++ and hardware interfaces to building LLM-powered pipelines and cross-platform apps.

At Qualcomm, I built automation systems adopted by production Physical Design teams. At Lovely Pets, I shipped a React Native CRM from scratch. My projects span visual question answering, real-time IoT tracking, and information retrieval.

Outside of engineering I've led 50+ volunteers at Rendezvous (IIT Delhi's cultural fest) and designed competitive programming problems for 1,000+ participants.

B.Tech — Electrical Engineering

Indian Institute of Technology Delhi

2022 – 2026 CGPA: 7.61

Class XII — HSC

Shri Shivaji Science College, Nagpur

2021 90%

Class X — SSC

Late Matadinlalji Jaiswal School, Nagpur

2019 93.6%

Scholastic Achievements

Academic Excellence

Consistently ranked among the top performers in national competitive examinations.

🏆

JEE Advanced

Rank 1484 — Top 0.6% among 250,000 students across India (2022)

🎯

JEE Mains

Top 0.73% out of 1.5 million candidates in the nationwide examination (2022)

📊

MHT-CET (PCM)

99.66 percentile — Top 0.34% out of 300,000 candidates (2022)

🥇

Certificate of Excellence

Rank 1 in Class X for outstanding academic performance & overall merit (2019)

Experience

Work Experience

Real-world impact through engineering internships at leading companies.

Interim Engineering Intern — Software

May 2025 – Jul 2025

Qualcomm

Built a modular automation script using TCL to detect, rank, and fix violations across physical-design workflows.
Created an LLM-driven query system converting natural-language questions into structured actions for fast analysis.
Delivered a production-ready system adopted across PD teams, improving efficiency and reducing fix times by 1/3rd.

Software Development Engineer Intern

May 2024 – Jul 2024

Lovely Pets LLP

Developed front-end modules for a cross-platform CRM using React Native, enabling efficient client management.
Designed and optimized MySQL database schemas to improve data handling, boosting shop productivity by 35%.
Contributed to end-to-end development of the app as part of a team, ensuring scalability for multi-store operations.

Projects

What I've Built

A mix of ML systems, embedded hardware, compilers, and full-stack applications.

Visual Question Answering using LLMs

Jul – Nov 2025

Prof. Sougata Mukherjea

Built a Visual QnA system with structured datasets supporting MCQ and open-ended queries. Engineered scalable pipelines for image cleanup, token parsing, inference decoding, and LLM fine-tuning. Developed a CLI, REST API & Gradio interface with real-time instrumentation.

Python LLM Fine-tuning Gradio REST API PyTorch

🔗 Links coming soon

NEXTBUS — Real-Time Smart Bus Tracking

Mar – Apr 2025

Prof. Subrat Kar

Designed a real-time bus monitoring solution with GPS-enabled ESP32 tracking units. Built data pipelines for signal collection, ETA modeling, bilingual display, and optimized communication. Achieved accurate ETAs for 500+ commuters, reducing latency by 20%.

ESP32 GPS IoT C++ ETA Modeling

🔗 Links coming soon

Personalized Offer Ranking from Transactional Data

Nov – Dec 2024

Designed a solution to personalize offers by analyzing user transactions and extracting relevance signals. Built a training pipeline using LambdaMART & XGBoostRanker with NDCG, MAP@7 scoring, and group-aware SMOTE. Improved ranking accuracy by 32% through tuned hyperparameters and top-K reranking.

Python XGBoost LambdaMART SMOTE Ranking

🔗 Links coming soon

Advanced Search Engine & Corpus Q&A Tool

Oct – Nov 2023

Prof. Rahul Garg

Designed a search system to extract information from large text corpora for natural-language queries. Built retrieval pipelines using hash maps, tries, scoring functions, and max-heap ranking to return top-K results. Enhanced answer quality by providing high-relevance context to LLMs.

C++ Trie Heap IR LLM

🔗 Links coming soon

E++ Expression Evaluator & Compiler

Aug – Sep 2023

Prof. Subodh Sharma

Built a compiler-like system to parse & evaluate expressions, implementing a custom large-integer class using OOP. Expression parsing used a binary tree & stack-based model with AVL symbol tables for fast lookups. Improved execution reliability through efficient memory management.

C++ OOP AVL Tree Compiler Design DSA

🔗 Links coming soon