Mar 2024 – May 2025 · 1 yr 2 mos · San Francisco, California, United States · Remote

• Architected the product backend using FastAPI, implementing critical features like chat streaming operations, file management via OneDrive, and role management modules, ensuring seamless AI-based interactions.
• Pioneered the development of a multiclass format reader and meta filter for MongoDBVectorStore, addressing specific AI business needs and enabling advanced query handling within Llama-Index.
• Generated synthetic datasets for AI training, aggregating over 100,000 conversations from platforms like StackOverflow, HuggingFace, and Kaggle, to enhance the AI models' accuracy and relevance in the customer support domain.
• Integrated Cross Encoders with Generative AI technologies, leveraging Langchain, Hugging Face, and Llama to refine semantic search results, significantly boosting the chatbot's response precision and user experience.

May 2023 – May 2025 · 2 yrs · Gurugram, Haryana, India · On-site

• AI-Powered CRM Productivity Application:
• Developed an AI-powered productivity application for CRM systems like Zendesk and Salesforce, crafting over 100+ FastAPI routes to support both admin and agent functionalities.
• Engineered advanced search features, including Vector Search and Keyword-based search, which is based on dimensional embedding, to deliver highly precise relevant ticket recommendations.
• Built a GEN AI-powered conversational bot utilizing a RAG pipeline, integrating technologies such as Langchain, Hugging Face, and Llama to enhance query accuracy and relevance.
• Enhanced search capabilities by integrating Cross Encoders for re-ranking semantic search results, improving accuracy and relevance for the Generative AI chatbot.
• Designed and implemented a robust data ingestion pipeline using RabbitMQ, incorporating data transformation by OpenAI Generative model and BERT-based topic modeling, with MongoDB for storage.
• Optimized system performance by leveraging event loops for parallel processing, utilized Celery for background task and Redis for caching to boost system responsiveness and reduce load times.
• Implemented secure authentication mechanisms using JWT-based authentication, OAuth2.0 authorization code flow, and OAuth2.0 OpenID Connect flow.
• ======================================
• Automated Attendance and Salary System:
• Streamlined HR operations by designing a Python-based cron job to automate attendance processing, reducing a week-long manual task to seconds by efficiently fetching and processing data from Teramind and Greythr APIs in real-time.
• Engineered and deployed high-performance RESTful APIs using FastAPI for calculating employee salaries, seamlessly integrating with Dataverse for data management, RabbitMQ for robust asynchronous processing, and the ICICI bank portal for secure financial transactions.
• Leveraged Docker for containerization, ensuring consistent and scalable deployments across all environments.

Aug 2022 – Oct 2022 · 2 mos

• This Internship is under IBM.

Oct 2021 – Nov 2021 · 1 mo