Sandeep GSC

Generative question and answers by Parsing BOSCH Domain documents with LLMs through LangChain and Lllama index with Chainlit UI using Generative AI. Working with the latest LLMs like llama, mpt, longchat, StableBeluga. Experience in solving problems that involve integrating NLP techniques using the Transformers library. (Text Classification, Multilingual Named Entity Recognition, Text Generation, Summarization etc.). also training the models with Custom Data. Experience on prompt engineering. Worked with chain of thoughts, ReAct(Reason+Action) in Langchain. Worked on generative QA using Retrieval-augmented generation (RAG). Trained RAG and HF Embedding models. Produced extractive and generative QA using Haystack. Worked on Image Segmentation for copper corrosion detection Using Fastai.

Predicting Lungs Disease using Deep Learning, model used VGG16, with an accuracy with 96%. Recommendation system for a hotel chain on food items. Experienced finetuning techniques to suit custom requirements. Data wrangling and selecting best features to develop better ML models.