About
Computer Science student specializing in AI/ML and building full stack production systems. Architected transformer-based computer vision models achieving 95% accuracy and engineered scalable RAG systems processing 100K+ vectors. Delivered 15+ production websites and AI-powered applications generating measurable business impact.
15+
Production Websites
100K+
Vectors Processed
95%
Model Accuracy
Projects
Agritech Disease Detection & Soil Analysis Platform
ReactNode.jsFastAPIGemini 2.0PineconeMongoDB
- ▸Architected cross-platform agri-AI system with Gemini 2.0 Flash RAG and Pinecone vector DB, enabling 3-second crop disease diagnosis across English, Hindi, and Kannada
- ▸Engineered semantic search pipeline using Gemini-001 embeddings to query 100K+ disease vectors with 150ms latency, achieving 92% retrieval accuracy
- ▸Developed ML recommendation engine processing 8 soil parameters (NPK/pH/CO2) via FastAPI, delivering 85%+ accuracy in crop-soil matching
- ▸Built scalable backend with MongoDB JWT authentication, UploadThing CDN integration, and unified Express API gateway serving 1000+ concurrent users
FormFit - Transformer-Based Workout Form Analyzer
PyTorchMediaPipeFlaskTensorFlow
- ▸Engineered PoseFormer transformer architecture achieving 95% accuracy at 33 FPS, outperforming CNN/RNN baselines by 15% with 8x faster inference (18ms vs 144ms)
- ▸Automated video-to-keypoint preprocessing pipeline to build 3,000+ sequence skeletal database, reducing data requirements by 60% through pose normalization and 2x augmentation
- ▸Developed real-time correction system using 3D keypoint smoothing and TTS voice feedback, providing professional-grade form analysis at zero marginal cost
- ▸Integrated Gemini API for personalized AI diet planning with SQL persistence, optimizing model convergence in 15-25 epochs
MERNReact NativeComputer Vision
- ▸MERN E-commerce Platform: Full-stack marketplace with Stripe integration, 10K+ monthly users
- ▸React Native Fitness Tracker: Cross-platform app with offline-first architecture, 5K+ downloads
- ▸Computer Vision Labeling Tool: Automated data annotation pipeline reducing manual work by 70%
Experience
Freelance Full-Stack & AI Engineer
Remote
2022 - Present
- Built AI-driven content automation pipelines for international NGO, reducing content production time by 60%
- Integrated Razorpay, Stripe, and third-party APIs for 5+ online education platforms, processing ₹50L+ in transactions
- Delivered 15+ production-ready websites for US, EU, and Indian clients (fortimark.co, rajadhaniholidays.in, kaizenworkspace.com)
- Maintained 98% client satisfaction rate with average project turnaround of 3-4 weeks
Featured Websites
Recognition
Lead Speaker - Startup Mahakumbh 2025
National Level
Presented AI powered agritech software solution as Top 7 finalist to Hon'ble Commerce and Industry Minister of India.
Education
Bachelor of Engineering in Computer Science
Global Academy of Technology, Bengaluru
Expected May 2026
CGPA: 6.89/10 (6th Semester)
Relevant Coursework
Machine LearningData ScienceDBMSSoftware Engineering
Skills
Languages
PythonJavaScriptTypeScriptSQLC++
AI/ML
PyTorchTensorFlowTransformersComputer VisionRAG SystemsLangChain
Web
ReactNode.jsExpressFastAPINext.jsReact Native
Cloud & Tools
VercelAWSGCPDockerGitApache SparkMongoDBPostgreSQL
Specialized
Vector Databases (Pinecone, Chroma)API IntegrationPayment Gateways (Razorpay, Stripe)
Languages
English
Fluent
Hindi
Fluent
Kannada
Fluent
Contact
Ready to build something amazing? Let's discuss your next project.
vgbofficial333@gmail.com
Phone
+91 9663873834
GitHub
github.com/updownvizzii
justanothervizzard
© 2026 Vignesh Ganaraja Bhat. All rights reserved.