Building intelligent systems that actually work in production
I'm a final-year B.Tech AI/ML student who believes in writing code that doesn't just work, but works well. Currently turning caffeine into production-ready AI systems and occasionally solving LeetCode problems to prove I remember what algorithms are.
- 🔬 Research Intern @ IIIT Nagpur - Teaching machines to spot skin lesions (98.65% accuracy, not bad for artificial eyes!)
- 🚀 Full-Stack AI Developer - Building systems that serve 100+ users without catching fire
- 📊 CGPA: 8.48/10 - Because perfect scores are overrated, but good enough scores get you places
- 🏆 Problem Solver - 100+ LeetCode problems conquered (and counting)
The system that makes HR's life easier (and mine more complicated)
Tech: Flask, Python, Qdrant, LangChain, Docker
- Built a production-grade ATS that doesn't crash when 1000+ users hit it simultaneously
- Reduced resume screening time by 60% (HR teams everywhere rejoice)
- 95%+ parsing accuracy because "close enough" isn't good enough
- Real impact: Currently handling actual recruitment workflows
Teaching computers to spot industrial flaws better than humans
Tech: PyTorch, YOLOv8n, Tesla P100, Roboflow
- Trained on 2,760 images of metallic surfaces (yes, I looked at that many metal photos)
- GPU-accelerated training because waiting is for the patient
- Published: Available on Hugging Face for other researchers
Making AI speak my mother tongue
Tech: SpeechT5, HuggingFace, HiFi-GAN
- Fine-tuned SpeechT5 for Telugu using 8,576 audio samples
- 116+ downloads on Hugging Face (people actually use this!)
- Built custom transliteration pipeline (because Telugu script is beautiful but complex)
C++ project that processes logs faster than you can generate them
Tech: Modern C++17, Threading, STL
- Custom thread pool architecture with lock-free operations
- Handles millions of log entries without breaking a sweat
- Use case: Perfect for high-frequency trading systems (if I ever go to Wall Street)
- 🥇 Code Spark Competition 2024 - 1st Place (HackerEarth platform)
- 🥇 Data Hunt Hackathon 2023 - 1st Place (Analyzed 75K+ airport records, helped solve a digital theft case)
- 🔬 Research Intern @ IIIT Nagpur - Working on skin disease detection (98.65% sensitivity achieved)
- 📜 AWS Certified AI Practitioner - Because cloud skills pay the bills
- 🎓 CGPA: 8.48/10 - Consistent performance without burning out
- 🔬 Research: Improving skin lesion classification using ConvNeXt and ViT+CBAM hybrids
- 💼 Building: Production-ready AI systems that don't fall over under load
- 🧠 Learning: Advanced system design and LLM optimization techniques
- 📚 Teaching: Mentoring junior students in AI/ML concepts at AI HUB
"Code should be written for humans to read, and only incidentally for machines to execute."
I believe in building systems that are:
- Reliable - Works in production, not just on your laptop
- Scalable - Handles growth without architectural rewrites
- Maintainable - Future you (and your team) will thank you
I'm always interested in discussing:
- AI/ML system architecture
- Production deployment challenges
- Research collaboration opportunities
- Why Python is simultaneously the best and worst language ever created
Best way to reach me: Drop an email at [email protected]
Response time: Usually within 24 hours (unless I'm debugging something particularly stubborn)