00. Welcome

Hi, I'm Vimansh Mahajan

< Software Engineer />

Computer Science Engineering student at IIIT-Delhi with expertise in Full Stack Development, Machine Learning, and Natural Language Processing

Vimansh Mahajan
Fortune 500
SWE Intern @ BNY
Researcher
@ IIIT Delhi
Leadership
TEDx | Senate | Placement
Get In Touch View Projects

01. About Me

I'm a Computer Science Engineering student at IIIT-Delhi, specializing in building scalable full-stack applications and developing machine learning solutions for real-world problems.

My experience spans from software engineering at Bank of New York to research in NLP and computational chemistry. I'm passionate about leveraging technology to create impactful solutions and constantly exploring new technologies to solve complex challenges.

Dean's List Award

IIITD Academics (2024-25)

Flipkart GRiD 7.0

National Semi-Finalist (2024)

Adobe GenSolve

Top 5% Finalist (2024)

02. Education

Indraprastha Institute of Information Technology, Delhi (IIIT-Delhi)

2022 - Ongoing

Bachelor of Technology (B.Tech.) in Computer Science Engineering

  • Awards: IIITD Dean's List Award Winner: Academics (2024-25)
  • Achievements: Top 5% Finalist of Adobe GenSolve Hackathon 2024
  • Leadership: Organiser @ TEDxIIITD, Member @ Placement Cell IIITD, Student Senate Representative
  • Entrance: JEE Advanced 2022 Qualified (AIR: 10,109), JEE Mains 2022 99 %ile (AIR: 9,200)

03. Professional Experience

Software Engineer Intern

Bank of New York (BNY)

May 2025 - July 2025
  • Built and deployed a content automation platform using Angular, Spring Boot, Oracle DB, and Python-integrated LLM
  • Transformed rough finance notes (text/images) into polished articles by identifying domain-specific segments and contextually placing visuals
  • Managed containerization and deployment on Google App Engine
Angular Spring Boot Oracle DB Python LLM GCP

Undergraduate Researcher

Complex Systems (CoSy) Lab

Jan 2025 - May 2025
  • Built RecipeGPT, a recipe generation system using the Recipe1M dataset, achieving human-like outputs with BERTScore > 0.85 and ROUGE score of 0.32
  • Applied a custom NER model to extract structured ingredient entities, enhancing input consistency
  • Developed a 3-stage GPT-2 fine-tuning pipeline: pretrained on 100k+ recipe instructions, conditioned on raw ingredients and titles, then fine-tuned with NER-enhanced inputs
GPT-2 NER PyTorch NLP Transformers

Undergraduate Researcher

TavLab

Aug 2024 - Dec 2024
  • Predicted IC50 values for HDAC6 inhibitors using ADMET and BlueDesc datasets by extracting molecular descriptors
  • Built and optimized ML models with advanced data preprocessing, feature selection, and statistical analysis
  • Accelerated compound screening and aided potential drug candidate identification through computational chemistry
Machine Learning Python Bioinformatics Data Analysis

04. Featured Projects

01

OptiWealth: Portfolio Management & Analytics

Nov 2025

Built a comprehensive portfolio management platform with Spring Boot backend and Python microservices delivering advanced financial analytics, risk diagnostics, and AI-powered insights.

  • Implemented VaR/CVaR risk metrics, ARIMA/GARCH forecasting, and Monte Carlo simulations
  • Built efficient frontier optimization with Sharpe ratio maximization
  • Integrated real-time market data via yFinance with JWT-secured REST APIs
Spring Boot Flask Python PostgreSQL NumPy Pandas JWT
View Code
02

Healthcare Answer Summarisation

Apr 2025

Built a BART-large-CNN + LoRA pipeline to generate perspective-specific summaries from medical Q&A data. Outperformed FlanT5 baseline by +3.24 BLEU and +0.0355 BERTScore.

  • Designed dual-head classifier for perspective classification
  • Implemented selective fine-tuning on hard examples (BERTScore < 0.84)
  • Achieved BLEU 5.61, BERTScore 0.8782
Python PyTorch BART LoRA Transformers
View Code
03

OncoHelp AI: Medical Decision Support

Dec 2025

AI-assisted medical decision support system for brain MRI triage integrating computer vision models, explainability techniques, uncertainty estimation, and multimodal LLMs for structured medical insights.

  • Built multi-model pipeline with ResNet18 and EfficientNet-B1 for brain tumor classification
  • Implemented Grad-CAM visualizations and Monte Carlo Dropout for uncertainty quantification
  • Integrated Gemini 2.5 Flash multimodal LLM for cautious medical summaries with scoped Q&A
PyTorch TensorFlow Streamlit Gemini LLM Computer Vision Grad-CAM
View Code

05. Technical Skills

Programming Languages

C++
Python
Java
JavaScript

Frameworks & Libraries

Spring Boot Node.js Express.js Angular Bootstrap PyTorch Hugging Face

Databases & Tools

PostgreSQL MySQL Oracle DB Git JUnit REST APIs

Specializations

Machine Learning Natural Language Processing Artificial Intelligence Data Science Full Stack Development Data Structures & Algorithms System Design Computer Networks DBMS

Certifications

Quant Finance & Algo Trading Ethical Hacking Python Bootcamp

Competitive Programming

Codeforces LeetCode

06. Get In Touch

I'm currently open to new opportunities and collaborations. Feel free to reach out!

Email

vimansh22572@iiitd.ac.in

LinkedIn

Connect with me

GitHub

Check out my code