Work & Internship

I have worked with many evolving startups as ML and DL Developer, Designer and Software Architect. I have also worked with some well established companies mostly as AI Developer. I love organising events and that is why I am also involved with many opensource communities as a representative.

Python with AIML

  • Built AI chatbots using LangChain and Gemini, leading to a 40% boost in user engagement.

  • Developed custom image/video generation tools with Stable Diffusion, enhancing visual output quality.

  • Automated tax calculations through a chatbot, cutting manual queries by 50%.

  • Improved backend performance with scalable APIs using FastAPI and Django.

  • Streamlined workflows via Docker and Make.com, increasing overall team efficiency by 35%.

Differenz system
a computer chip with the letter a on top of it
a computer chip with the letter a on top of it

Jul 2024 - Present

Python with AI

  • Spearheaded enhancements to the Drive Inspector project by incorporating advanced Number Plate Detection, OCR, and RTO Database Integration functionalities, leading to a 30% reduction in manual data entry tasks and a 20% improvement in real-time compliance notifications.

  • Improved Indian number plate detection accuracy to 90% by fine-tuning and adding accurately annotated data.

  • Orchestrated the capture and processing of roadside video footage, resulting in the creation of robust custom datasets that enhanced model training effectiveness by 40%.

  • Created an admin dashboard to monitor vehicle detection and details, which increased decision-making by 25%.

Axisray
person sitting front of laptop
person sitting front of laptop

Jan 2024 - Apr 2024 • 4 mon

AI Intern

  • Actively contributed to diverse AI projects, collaborated in an 8-member team.Implemented AI person counting system with 91% accuracy and number plate recognition project having 92% accuracy.

  • Created AI fire and smoke detection system with an 85% accuracy rate for safety-critical applications. Implemented an on-site AI-driven system to detect Personal Protective Equipment (PPE) with an impressive accuracy rate of 88%. Prepared 80 computer vision datasets, ensuring data integrity, proper labeling, and conversion to YOLO TXT format.

Tusker AI
white robot toy holding black tablet
white robot toy holding black tablet

Aug 2022 - Mar 2023 • 8 mon