Developed and deployed a high-accuracy classification model to identify over 30 different types of gas cylinders used in automated gas filling robots. The model achieved 98% accuracy and was optimized for real-time inference on NVIDIA Jetson hardware, enabling seamless edge deployment with minimal latency. This AI solution significantly improved automation reliability and reduced manual verification steps.
Designed and deployed an AI-based nut sorting model capable of accurately classifying and sorting various nut types with 99% accuracy and precision. Leveraging computer vision and deep learning, the system was tailored for real-time performance, reducing misclassification and significantly improving sorting efficiency. This solution is optimized for industrial deployment, ensuring reliability in high-speed manufacturing environments.
The Multi AI Agent System is an intelligent, autonomous framework designed to automate the financial record-keeping process. CrewAI, OpenAI API, Document AI API are used to build this Multi AI Agent system consist of: 1) Data Ingestion AI Agent 2) Categorization AI Agent 3) Reconciliation AI Agent 4) Report Generator AI Agent 5) Query AI Agent This increase the efficiency of my client's team by around 60% and revenue get approximately 2x
Developed a fully automated AI agent-powered marketing workflow using n8n to streamline content creation across multiple channels. 1) AI-driven blog post generation with custom image generation. 2) Automated LinkedIn post creation, with relevant AI-generated visuals. 3) A standalone AI image generator with OpenAI. 4) Image search functionality to fetch past generated images. This increase the efficiency of my client's team by 60-70% and revenue get 3x since they now can land more clients




