Deputy Manager - AI Engineer (Lead) at Utopia Deals (2025-11 – Present)
Working on Amazon advertising efficiency and agentic approaches to campaign management.
- Working on Amazon advertising efficiency by reducing cost per sale and introducing an agentic approach to managing campaigns.
- Provided clear insights and reports to help Utopia's campaign managers work more efficiently and reduce manual effort.
Sr. Machine Learning Engineer Gen. AI at Drox Labs (2024-06 – 2025-11)
Built and deployed generative AI systems using OpenAI, LLaMA, and Gemini models for real-time interaction and content creation.
- Built and deployed generative AI systems using OpenAI, LLaMA, and Gemini models for real-time interaction and content creation.
- Developed AKO, an image-to-video generator with character animation from text prompts using vision and LLM models.
- Led the development of SONA, an LLM-powered chatbot for live market insights in Web3, integrated with Solana.
- Designed AKIRACORE, a smart contract generation assistant using LLMs to convert user inputs into deployable Solidity code.
- Developed scalable backend services using FastAPI for AI-driven applications and chatbot integrations.
Sr. ML Engineer Consultant (Freelance) at Bytecorp (2025-06 – 2025-11)
Built real-time vision systems for garbage bin detection and ALPR-based analytics.
- Built real time vision systems for detecting garbage bin status on moving trucks using optimized deep learning models.
- Delivered an ALPR based showroom analytics platform for vehicle recognition, people tracking, and footfall insights.
- Integrated live camera streams with dashboards to provide actionable data for operational and customer behavior analysis.
Sr. Machine Learning Engineer Computer Vision at Folio3 (2022-05 – 2024-06)
Led development of advanced vision systems including multi-view vehicle re-identification and autonomous tennis umpiring.
- Led development of advanced vision systems including multi view vehicle re identification, sports activity recognition, and an autonomous tennis umpiring solution.
- Optimized models with ONNX, TFLite, and TensorRT for efficient edge deployment and delivered production grade ML services with complete pipelines.
- Mentored junior engineers through internal AI apprenticeship sessions focused on practical machine learning and vision methods.
Research Engineer, Computer Vision at Retrocausal.AI (2021-05 – 2022-05)
Developed object detection and action recognition models for industrial workflows.
- Developed object detection and action recognition models for industrial workflows, supported by custom preprocessing and end-to-end training pipelines.
- Delivered a real time visual task monitoring system and deployed models through Django APIs for seamless product integration.
