🌟 I'm an Information Science & Engineering undergrad at RV College of Engineering, Bengaluru, passionate about building AI-powered products end-to-end — from ML pipelines to production infrastructure.

• 🎓 B.E. in Information Science and Engineering, RVCE (CGPA: 8.03) — Class of 2027
• 🏆 Smart India Hackathon 2025 Finalist for Sanjaya, a real-time sign language interpretation system
• 🔭 Currently building AI education tools, RAG-based research assistants, and computer vision systems
• 🌱 Exploring Mobile Development, Applied Machine Learning, and Agentic AI
• 💻 500+ problems solved on LeetCode, rated 1600+
• 📫 Reach me at shrinidhiachar857@gmail.com
• ⚡ Fun fact: I enjoy turning ideas into reality through code!

FastAPI React TypeScript Gemini Manim PostgreSQL AWS (EC2, RDS, S3) Nginx FFmpeg gTTS

• Architected and deployed a full-stack AI platform on AWS (EC2, RDS, S3) with Nginx reverse proxy and HTTPS/SSL for custom-domain access.
• Built an 8-stage AI pipeline that turns natural-language math/CS prompts into narrated animated video lessons — lesson planning → scene decomposition → Manim code generation via Google Gemini.
• Designed a custom Animation DSL with a validation layer converting LLM-generated JSON into safe Manim scripts, preventing rendering failures before compilation.
• Built an audio-video sync pipeline (gTTS + FFmpeg) and a React/TypeScript studio backed by FastAPI with JWT/OAuth auth and S3 asset storage.

Flask LangChain ChromaDB Gemini API PostgreSQL React TypeScript TailwindCSS

• Engineered a ReAct-based LLM agent with a RAG pipeline over ChromaDB, using Google's text-embedding-004 model for document-grounded Q&A and reference-aware paper drafting.
• Integrated Gemini 3.5 Flash to autonomously read cited papers, inject references, and generate/refine LaTeX content — significantly cutting manual writing effort.
• Architected a Flask REST backend + React/TypeScript frontend for a seamless researcher workflow from ingestion to draft publication.

Python OpenCV YOLOv8 Hough Transform NumPy

• Built a computer vision pipeline to digitize chalkboard content from lecture videos in real time, reducing bandwidth while preserving educational content.
• Combined HSV masking, Canny edge detection, and Hough line validation with YOLOv8 segmentation to eliminate instructor occlusion.
• Designed a Temporal Confidence Voting (TCV) algorithm achieving 97.41% precision at >25 FPS.

Raspberry Pi 5 Python MediaPipe Transformers Flutter Bluetooth

• Built an end-to-end sign language interpretation system combining computer vision, transformer-based gesture recognition, and text-to-speech.
• Engineered a low-latency, on-device inference pipeline using MediaPipe landmark extraction and transformer models for dynamic gesture classification.
• Implemented Bluetooth communication between a Raspberry Pi 5 and a Flutter app for real-time wireless speech synthesis and portable assistive communication.

• 🏆 Smart India Hackathon 2025 Finalist
• 💻 LeetCode — 1600+ rating, 500+ problems solved
• 📜 AI for Economics — Certified
• 📜 Data Science — Certified
• 📜 C++ DSA — Certified
• 📜 Salesforce Agentforce — Certified

📊 GitHub Stats

👀 Profile Views