Junior Data Engineer
CurrentServed as a key member of a dedicated Data Science team responsible for developing a specialized generative Chat-GPT style model tailored for an exclusive Discord channel utilized by users. Collaborate with a team to develop a user query answering system utilizing the Mistral 7b LLM with RAG, enhancing customer engagement and support capabilitiesDesign and implement data scraping and preprocessing routines from Discord channels, ensuring robust dataset creation for model… Show more Served as a key member of a dedicated Data Science team responsible for developing a specialized generative Chat-GPT style model tailored for an exclusive Discord channel utilized by users. Collaborate with a team to develop a user query answering system utilizing the Mistral 7b LLM with RAG, enhancing customer engagement and support capabilitiesDesign and implement data scraping and preprocessing routines from Discord channels, ensuring robust dataset creation for model training.Integrate LangChain with the LLM to streamline API interactions and data handling, increasing system responsiveness and accuracy.Optimized model performance by implementing Flash Attention techniques, significantly accelerating response times while maintaining output quality.Achieved a drastic reduction in GPU RAM requirements from 14GB to 3.5GB by configuring model quantization, enhancing scalability and reducing operational costs. Show less
