Location: Navi Mumbai, India
Education: Final Year B.Tech in Electronics & Telecommunication (Minor in AI) – RAIT
CGPA: 8.79 / 10
Email: ojasshen1584@gmail.com

Passionate about building intelligent systems that merge AI, Data, and Automation.
Experienced in LLMs, RAG, Computer Vision, NLP, and Time-Series Analytics, focusing on real-world GenAI applications and data-driven reasoning systems.

Swift Multi-Modal Assistive Translator

• macOS Vision + ML Model + Haptic Glove translating speech-to-Braille for deafblind accessibility.
• Real-time transcription and haptic feedback using a 6-motor glove system.
• Decoupled Vision and Haptic modules for independent debugging and deployment.
  Stack: Swift · CoreML · BiLSTM · AVFoundation · Python

Clinical Decision Support System (RAG)

• RAG pipeline sourcing NCCN & ICMR guidelines, improving relevance by 40%.
• Processed 500+ complex PDFs into Weaviate vector DB via Python + Marker.
• Benchmarked Llama 3.1, Mistral; selected Llama 3.1 for best reasoning & faithfulness.
  Stack: Python · LangChain · Weaviate · Llama 3.1 · NLP

Marker-Based UAV Navigation (Patent Filed)

• Robust indoor autonomy using ArUco markers.
• Custom ID-based logic boosted detection accuracy by 30%.
• Adaptive motion planning reduced path error by 40%.
  Stack: Python · OpenCV · ROS · Computer Vision

Leaf Disease Detection (YOLOv8)

• 97% accuracy in classifying plant diseases.
• 25% robustness improvement under variable lighting via augmentation.
  Stack: Python · YOLOv8 · Deep Learning

A Star Path Planning for Drones

• Optimized A* in ROS/Gazebo; 95% obstacle avoidance.
• Reduced mission time by 15% through re-planning.
  Stack: Python · ROS · Gazebo

“Turning complex data into meaningful intelligence that powers human-like systems.”

Last Updated: October 2025