Looking for a job

Tisha Jindal

@tishajindal

Message

I build secure full-stack and machine learning applications from end to end.

India

Message

What I'm looking for

I'm looking for opportunities to build secure, scalable, user-focused applications while growing across full-stack development, machine learning, and software engineering.

I've built PR Guardian, a deployed GitHub App that analyzes pull request changes with the Gemini API and posts automated code-review feedback directly on GitHub.

I implemented HMAC-SHA256 webhook verification and GitHub App installation authentication with Octokit, replacing long-lived personal access tokens. I also engineered the pipeline for filtering changed files, chunking large diffs, and triggering reviews on pull request events.

For Atom Forge, I built a molecular-generation deep learning pipeline using GNNs trained on ZINC-250K and an LSTM model for predicting seven molecular properties. I processed compounds with RDKit and deployed the full-stack application with FastAPI, Next.js, Railway, Vercel, and Docker.

I've also trained credit-risk models on 38K+ records, reaching 91% test accuracy and 0.87 ROC-AUC. I'm pursuing a B.Tech in Computer Science Engineering and apply software development lifecycle principles to secure, scalable, user-focused applications.

Experience

Work history, roles, and key accomplishments

PR Guardian Developer

GitHub

Jun 2026 - Aug 2026 (2 months)

Built and deployed a GitHub App that analyzes pull request code changes using Gemini API and posts automated code review feedback directly on GitHub. Implemented HMAC-SHA256 webhook verification and GitHub App installation authentication with Octokit.

NodeJS GitHub Apps Octokit Gemini API HMAC Webhooks REST APIs

Atom Forge Developer

Live

Apr 2026 - May 2026 (1 month)

Built an end-to-end deep learning pipeline for molecular generation using GNNs trained on ZINC-250K and deployed a full-stack app with FastAPI backend and Next.js frontend. Processed and validated compounds using RDKit for SMILES manipulation and descriptor extraction.

Python TensorFlow fastAPI GNN LSTM RDKit Docker Next.js Railway Vercel

Home Credit Default Risk Modeler

Home Credit

Sep 2025 - Oct 2025 (1 month)

Trained Logistic Regression and Random Forest models on 38K+ records with 80+ engineered features, achieving 91% test accuracy and 0.87 ROC-AUC. Applied SMOTE for class imbalance and identified top predictors via feature importance analysis.

Python scikit learn Pandas Logistic Regression Random Forest SMOTE Feature Engineering