Pget- Data Scientist
Bengaluru, Karnataka, India
- Working in various projects & POC's in the field of AI focused on Computer Vision, Natural Language Processing & Generative AI- Implementing various research papers in the field of AI to provide solutions for problems faced in the Aftersales of Mercedes Benz.- Next Gen BMS (Battery Management System) using AI Computer Vision:-Build a Object Detection model for detection Scratch & Dent , used Focal Loss + Soft Sampling and Few-Shot Learning to overcome class imbalance problem of the dataset .- Building an counting application using various model like CLIP-Count, ZSC, VL-Counter, Count-Anything and T-Rex (text and visual zero shot models). NLP + Generative AI:- Build a Text Classifier by Finetuning + Instruction Tuning over FLAN-T5(Spam-T5,model by JPMorgan released in 2023, a highly accurate Few-Shot classifier).- Language Detection of Text- BERTopic Modelling of CSAT(Customer Satisfaction) Data from CAC of Mercedes-Benz- Prompt Engineering + Prompt Optimization for various NLP Task- Llama 2 70B based ChatBot using RAGAchievement:- Winner of AI Ideathon 2.0, an AI innovation competition held across MBRDI.- Developer of the Month (May, 2024) of entire Aftersales Department of MBRDI.