- Developed an intelligent chatbot platform using LangChain, RAG (Retrieval Augmented Generation), HuggingFace models, Langsmith, and OpenAI's GPT-based APIs. - Enhanced user engagement by leveraging natural language processing to provide real-time, personalized recommendations and support for web discovery.- Integrated RAG to improve response accuracy by retrieving relevant data from indexed web sources. - Developed and tuned LLM models to improve the accuracy of responses, leading to a 20% reduction in fallback scenarios and enhancing overall customer engagement.- Worked with product managers to identify pain points in user interactions, proposing technical solutions that aligned with business goals.

- Utilized LLMs with RAG to create highly directed fuzzers that increased fuzz coverage up to 80%.– Created 1:1 digital twins of linux-native High Performance Computing workloads in clusters using C, Linux, and SLURM API.– Owned and configured the SLURM job scheduler to meet the needs of an NSF-funded research project.– Maintained and updated the software stack, including OpenMPI, Intel OneAPI, and custom package installations using Chef.– Modified SLURM code and configured features for advanced vulnerability detection.– Maintained and optimized a virtual research cluster with 200+ compute nodes.– Monitored cluster resources including compute nodes and storage using Zabbix and Grafana.– Created simulations of SLURM workloads to fuzz test workflows; and node, partition, scheduler, and logging configurations.– Analyzed and triaged HPC systems and network security oriented CVEs (vulnerabilities) with digital twin.– Automated SLURM installation and configuration management through Python and Bash scripts.– Enhanced the efficiency of job scheduling on the HPC cluster by refining SLURM configurations, resulting in a 30% reduction in job wait times and improved resource utilization.

Course: Computer Systems.

Course: Industry Forum.

- Developed a multi-channel conversational chat bot, leading to a 30% increase in usage.- Developed a customer portal using React for the front-end, Node.js for the back-end, and Web API for seamless communication between the client and server.- Built responsive, dynamic user interfaces in React, ensuring smooth navigation and user experience across various devices.- Developed scalable RESTful Web APIs in Node.js to handle customer data, authentication, and real-time updates, ensuring high performance and security.- Integrated the portal with backend systems for data storage, using JWT authentication for secure user access.- Visualized chatbot performance metrics through custom dashboards in Tableau through telemetry data from Task Execution Logs, with detailed KPI reports, enabling data-driven decision-making.- Optimized conversational flows, leading to a 20% increase in user satisfaction.

Won two awards for exceptional performance.- Conceived an speech-enabled chatbot for customer service automation using Azure’s Natural Language Processing stack (Azure AI Services, Azure Bot Service, Speech Service, and Language Understanding (CLU/LUIS)) for processing customer queries to answer customer questions in AI-generated speech in natural language. Generated >$100,000k in cost savings.- Innovated a tool to generate documentation of ETL pipelines in MS-SQL ETL (SSIS) packages to detail the program flow in pseudocode, saving ~$30,000 in development hours.- Spearheaded a web app for paperless statements sign-ups leading to +30% sign ups in 2 weeks.- Restructured and refactored legacy monolith applications to cloud-native serverless applications with microservices improving performance by 70%, with performance tuning and unit testing. - Reduced ETL job failures by 70% by identifying and fixing bad SQL queries in SSIS.- Automated building, testing, and publishing of applications with Azure DevOps Server, Github Actions and Octopus Deploy for source control, configuration management, and CI/CD pipelines.- Guided 3 junior team members to improve SLA metrics of service quality and business continuity by 15%, and an 80% reduction in code quality and deployment issues. Organized technical workshops.- Integrated Kafka to track real-time user interactions across multiple web applications, enabling enhanced data monitoring and analysis.- Analyzed bot telemetry data using custom queries in Azure Monitor Log Analytics to identify patterns in customer interactions and areas for improvement in conversational flows, contributing to a 15% increase in customer satisfaction scores.- Created Power BI dashboards to visualize performance metrics from NLU models and customer interaction data, leveraging business insights and feedback to improve response accuracy.