╔══════════════════════════════════════════════════════════╗
║ FROM RAW DATA → INTELLIGENT DECISIONS → PRODUCTION AI ║
╚══════════════════════════════════════════════════════════╝
3rd Year B.Tech AI/ML · JSSATE Bengaluru · CGPA 9.15 · 8+ Hackathons · 5 Deployed Systems
akash_g_patil = {
"role" : "AI/ML Engineer in the Making",
"researching" : "HalluciSense — LLM Hallucination Detection",
"target" : "Elsevier / Frontiers in AI Publication · 2025",
"deployed" : "GreenWave AI → greenwave-ai.vercel.app",
"cgpa" : 9.15,
"hackathons" : "8+ competitions · multiple wins",
"available" : True, # Open to internships & research
"location" : "Bengaluru, Karnataka, India",
}
# What separates me from other ML students:
print("I don't just train models.")
print("I engineer complete AI systems — from research to production.")Most students build for Kaggle. I build for the real world.
|
LLM Hallucination Detection Framework Research Project · Major Project JSSATE The problem: LLMs hallucinate. Confidently. Dangerously. The solution: A three-pillar hybrid architecture that detects, quantifies, and scores hallucinations in real-time. 94% detection accuracy 🎯 Target: Elsevier / Frontiers in AI · 2025
|
Emergency Vehicle Preemption System Build for Bengaluru 2.0 · AI/ML Track · Deployed The problem: Ambulances stuck at red lights. People die in those seconds. The solution: AI that clears every signal on the route — automatically, in milliseconds. Live in production → greenwave-ai.vercel.app
Built in 24 hours. Shipped. Featured in press. |
|
Predictive Strategy Engine ML system that thinks like an F1 race engineer. Live telemetry data → optimal pit stop windows → lap-by-lap strategy simulation.
|
Real-Time ML Pipeline Production-grade fraud detection on 284k transactions. Handles extreme class imbalance. Deployed with Docker.
|
| Layer | Technologies |
|---|---|
| AI/ML Core | PyTorch · TensorFlow · Scikit-learn · HuggingFace · OpenCV |
| LLM/NLP | RAG · Hallucination Detection · Prompt Engineering · FAISS · Transformers |
| Backend | FastAPI · Docker · WebSockets · REST APIs · Node.js |
| Frontend | Next.js · React · Tailwind CSS · JavaScript · HTML5 Canvas |
| Data | Pandas · NumPy · SQL · Streamlit · Matplotlib · Jupyter |
| DevOps | Git · Linux · Vercel · Cloud Computing · GitHub Actions |
| Metric | Value | Context |
|---|---|---|
| 🎓 CGPA | 9.15 | Top of class · JSSATE Bengaluru |
| 🏆 Hackathons | 8+ | Multiple wins · Build for Bengaluru 2.0 |
| 🚀 Live Systems | 2 | GreenWave AI + Portfolio |
| 🧠 ML Accuracy | 94% | HalluciSense detection |
| 📚 10th Grade | 95.8% | Kendriya Vidyalaya · CBSE |
| 📚 12th Grade | 92.5% | Alva's · PCMB |
2021 ── 10th · CBSE ████████████████████ 95.8% Kendriya Vidyalaya
2023 ── 12th · PCMB ██████████████████░░ 92.5% Alva's · Moodabidri
2023 ── B.Tech AI/ML ██████████████████░░ 9.15 JSSATE Bengaluru
└── Specialization: Artificial Intelligence & Machine Learning
8+ competitions · Build for Bengaluru 2.0 · Smart India Hackathon 2025
My personal winning formula — developed across 8 competitions:
✓ Idea clarity beats code volume (always)
✓ Deploy early — shipped > perfect (always)
✓ Invest 40% time in presentation (always)
✓ Build with familiar tools (always)
✓ Sleep is non-negotiable (always)
✓ Use AI tools ruthlessly — Cursor, GPT (always)
Built with: Next.js 14 · HTML5 Canvas · Framer Motion · TypeScript
120 AI-generated cinematic frames → scroll-linked animation
Scramble text reveal · Custom cursor · Noise grain overlay
Awwwards-level interactions · Built in 1 day · Shipped to production
I'm actively looking for AI/ML internships, research collaborations, and opportunities to build systems where accuracy genuinely matters.
| Channel | Link |
|---|---|
| 🌐 Portfolio | akashgpatil.vercel.app |
| akashgpatil23.05@gmail.com | |
| linkedin.com/in/akash-g-patil | |
| 📍 Location | Bengaluru, Karnataka, India |