Mahesh Deshwal

As a seasoned ML - AI Engineer, I leverage my expertise in end-to-end research, problem formulation, development, deployment, and presentation of scalable solutions in NLP, Computer Vision, Search & Recommendation using MLOps methodologies to deliver business goals. With a strong background in MLOps, Machine Learning, Deep Learning, Computer Vision, NLP, and Recommendation Systems, I have achieved notable tasks such as authoring a SOTA paper in 4 Benchmarks using Phi-3 for LLM evaluation and creating India's only open-source tool for personal financing and investment, NSE-Stock-Scanner.

Throughout my career, I have worked on various projects, including multimodel intent classification, text-to-image generation, and building the fastest LLM evaluator model. I have also fine-tuned embedding models, owned personalized recommendation systems, and built in-house generative AI solutions. My expertise in deploying recommendation engines, implementing auto question-answering pipelines, and designing named entity recognition systems has improved user engagement, reduced manual effort, and enhanced data quality.

I am passionate about staying updated with the latest tech and contributing to open-source projects, including Vector DB LanceDB. I have multiple official contributions with Google Research and Huawei Lab, and I am excited to continue making impactful contributions in the field of AI and ML.