Jan 2025 – Present · 1 yr 3 mos · On-site

• Developed an OCR-based information extraction system for financial documents using GenAI and document layout analysis, enabling accurate parsing of unstructured data.
• Built a GenAI-powered tool for automated unit test generation using LLMs and LangChain, supporting multiple languages including Node.js, React, TypeScript, and Python.
• Implemented context-aware prompting and retrieval techniques to enhance code understanding and accuracy of generated test cases.
• Leveraged state-of-the-art coding LLMs to generate robust, production-grade unit tests, reducing manual testing effort and improving code quality.

Jul 2024 – Jan 2025 · 6 mos · On-site

• Created a classification model by incorporating the Random Forest algorithm. Improved accuracy by integrating NLP-based methods, specifically utilizing XLNet model embeddings for enhanced performance.
• Implemented machine learning models to replace manual workflows, significantly improving speed and accuracy.
• Leveraged large language models from quantization to deploying them to optimize the workflow.
• Developed a multi-retrieval based RAG model to create QnA based assistance for business crucial documents.
• Applied advanced NLP techniques to improve existing classification models, resulting in scalable solutions and reduced operational costs.
• Worked on Generative AI approaches to be integrated with Finance workflow.
• Agentic AI : Developed a Chatbot using SQL agent in langchain incorporating it with Google Bigquery.
• Seamlessly deployed models by integrating CI/CD pipelines on Google Cloud Platform (GCP), ensuring efficient model lifecycle management with MLflow.