* Helped develop an ATS leveraging Large Language Models (LLMs) to enhance traditional ATS functionalities. Introduced features like detailed resume scoring, segmentation, and automated data collection, leading to significant improvements in data completeness and accuracy.* Engineered and deployed over 20 advanced prompt-based solutions using GPT-4o and Llama3, focusing on OCR-based resume scanning, scoring, and automated question generation, which greatly enhanced processing efficiency and accuracy.* Designed and implemented a robust data pipeline using AWS S3 and PostgreSQL, securely managing over 10,000 resumes and ensuring efficient storage and retrieval across multiple users.* Created 10 RESTful APIs using Python and Flask to enable seamless integration between various system components, facilitating a cohesive and responsive user experience.* Led benchmarking and optimization efforts for 5+ system modules, applying performance profiling and execution time analysis. Achieved a 60% improvement in run-time efficiency.* Conducted in-depth research on Automatic Speech Recognition (ASR) models, identifying that mid-level systems have higher Word Error Rates with strong accents compared to advanced systems like OpenAI Whisper enhancing accent detection in chatbot applications.* Implemented LLM guardrails by reviewing and applying insights from key literature, such as DecodingTrust, to evaluate models like GPT-4 and GPT-3.5 for toxicity, bias, adversarial robustness, privacy, and fairness.

At Fractal, I worked with the the retail division a Fortune 500 FMCG Company, as a Data Engineer and Data Scientist. Primary responsibilities:* Developed a SKU Simplification tool to optimize the Portfolio Rationalization process by creating a system of record for the portfolio’s performance by deriving insights from various data sources and providing recommendations for delisting opportunities to help cut down cost, saving the client ~ $150M in operational revenue * Built automated pipelines for data acquisition, processing and validation and CI/CD pipelines to deploy resources across different environments, reducing the time spent on these activities by over 30%* Lead all technical and functional activities for the US track of the project. Worked closely with various cross-functional stakeholders and business users to identify and implement enhancements to the tool based on business requirements, leading several client interactions. * Improved the quality of the dataset and the recommendation engine by increasing the coverage across all data lenses and introduced additional lenses metrics to provide users with more valuable insights and recommendations