Conducted research and development on multi-modal OCR systems, including Duckling, Donut, InternLm, and QwenVI, advancing document analysis and recognition technologies. Integrated NLP techniques to discern variations between documents, enhancing accuracy and efficiency, and aligning with cutting-edge projects in Generative AI. Streamlined development processes by implementing CI/CD pipelines, automating testing and deployment of AI services, and ensuring effective integration of ML solutions for cloud infrastructure and methodologies. Interacted with stakeholders to demonstrate product features, leading to successful production of deployment and ongoing management. Demonstrated expertise in AI/ML with hands-on experience in LLM models such as BLOOM, Imperia, Ollama, BERT, Claude Sonnet, and Mistral. Utilized API services like Open AI and Azure, and deployed agents including Open Hermes. Proficient in frameworks such as Crew AI and Functionary, contributing to strategic planning for innovative computer vision solutions.

Focused on preparing the dataset for training generation of spectrograms, extracting features of MFCCs, developing classification algorithms and a web interface to interact with the ML model. Utilizing Numpy, Tensorflow, sklearn, matlotlib and pythonQT.