class ML_Engineer:
def __init__(self):
self.name = "Dhananjay Sai Kumar K"
self.location = "Chennai, India"
self.role = "Machine Learning Engineer"
self.focus = ["Deep Learning", "Computer Vision", "Speech Systems", "Generative AI", "Robotics"]
self.philosophy = "From research papers → production-ready systems"
def build(self):
return "End-to-end ML pipelines that scale in the real world"-
Production-ready speech pipeline (Whisper + Demucs + Coqui TTS + Gemini API)
-
98% vocal separation clarity | ±0.15s lip-sync accuracy | 6× noise reduction
-
Processes 10-min videos in 5–8 minutes on GPU
-
Modular ETL pipeline: Download → ASR → Separation → NMT → TTS → Rendering
-
Real-time HCI system (OpenCV + MediaPipe + Computer Vision)
-
Features: Draw, Erase, Undo/Redo, Color switching, Multi-page canvas
-
Led algorithm design & ML-UI integration for 3-member team
-
Patent filed (2025) — Novel gesture-based drawing interface
-
Full-stack + AI tourism platform (Next.js + Prisma + GeoJSON + Embeddings)
-
<2s load time | 95+ Lighthouse scores
-
Interactive GeoJSON destination explorer
-
AI trip planner using embeddings, clustering & scoring
-
Satellite forecasting using DDPM + U-Net (INSAT-3DR)
-
Predicts future cloud frames
-
Custom multi-band spatiotemporal ETL pipeline
-
Optimized noise scheduling & architecture
-
Autonomous navigation system (YOLO + ROS2 + ZED2i)
-
92% detection accuracy | 97% interaction accuracy
-
Depth maps + terrain-aware obstacle avoidance
-
Robotics team collaboration
-
SAR → RGB conversion using U-Net
-
94.4% colorization accuracy
-
Optimized preprocessing & training pipeline
- Patent Filed (2025) — Gesture Craft Intelligent Drawing System
- Hackathon Finalist — Hack for Purpose
- 1st Prize — National Science Day Video Competition
- Class Representative (2024–25)
- Community Outreach — STEM Programs
- IBM Machine Learning Professional Certificate
- Deep Learning & Reinforcement Learning — IBM
- Supervised & Unsupervised ML — IBM
- Exploratory Data Analysis — IBM
- NPTEL: Python, DSA, Machine Learning
"Building AI that bridges research and reality — one deploy at a time."
