RUHAN SIDDIQUI

Developed LLM's Langchain and OpenAI App. Created a Knowledge base chatbot using Langchain and various open-source LLMs (e.g., LLAMA 7B, Mistral AI). Constructed embedding models for precise responses. Implemented storage of embeddings in Chroma and FAISS vector DB. Fine-tuned OpenAI models (text-davinci-003, gpt-3.5-turbo) for specific use cases. Applied LEFT LORA techniques to fine-tune LLM models using provided data. Built a language translation model. Developed OCR models for real-time image scraping. Implemented Text-to-Speech (TTS) using style TTS and Bark AI for multiple languages. Created a Speech-to-Text (STT) model using Whisper for Hindi, English, and Punjabi. Applied prompt engineering techniques for improved output. Developed a Resume parser using Mistral AI LLM models. Constructed a Machine Learning classification model for layer identification. Established a Retrieval Augmentation Generation (RAG) pipeline for fetching accurate data from source documents to enhance response generation.

Worked on Python and Advance Pandas. Worked on Selenium with Python. Worked on writing SQL queries and create Reports.

Extracted and transformed unstructured live data from Telematics devices using Python in the Data Pipeline. Developed structured format, created a pipeline to store data in MySQL DB. Utilized MySQL for data analysis and query execution based on client specifications. Collected real-time data from Caruso website via Postman API, ingested into MongoDB using Python. Performed aggregation operations to transition nested data to a non-nested format in MongoDB. Established a pipeline environment for processing and storing live data in MongoDB Database. Applied MongoDB queries for data analysis in line with client requirements. Selected impactful KPIs in Dashboard designer, reported daily work in Agile Environment using Jira for informed business decision-making.