• Build and deploy autonomous agent workflows in Python using uAgents, implementing orchestration, integration, testing, and debugging for execution through Agentverse and ASI:One.

• Lead a 7-intern effort on ASI:One UI/UX improvements, translating user pain points into feature proposals, Excalidraw flow maps, Figma wireframes, and coordinated sub-team execution across multiple feature tracks.

• Partner closely with my manager and senior leadership to shape architecture and workflow direction, debating design tradeoffs and refining technical decisions before implementation.

• Improve ASI:One usability and personality-agent behavior by turning new agent capabilities into clearer, more intuitive product-facing experiences.

• Demo agent workflows and feature improvements regularly, communicate their user value, and help promote ASI:One’s capabilities externally.

• Contributed to a redesign direction that improved user engagement by 10% over two weeks based on direct testing and feedback.

Grader for CS151 - Object Oriented Design

Grader for CS46B - Data Structure & Algorithms

Grader for ENGR10

• Developed and optimized backend services in Python and Java for production workflows supporting Bosch business operations, with a strong focus on performance, scalability, and clean service design.

• Worked extensively with AWS DynamoDB for backend data access patterns and service integration, helping reduce API response time by 30% during high-traffic production periods.

• Built and improved CI/CD pipelines using Jenkins and Docker, strengthening deployment reliability, reducing deployment errors by 50%, and accelerating release cycles by 40%.

• Developed data preprocessing pipelines for AI-driven analytics and forecasting use cases, transforming raw business data into usable inputs for downstream predictive systems.

• Partnered cross-functionally with product and analytics stakeholders to productionize forecasting insights, contributing to a 20% improvement in inventory forecasting accuracy and a 15% reduction in excess inventory storage costs.

• Worked in environments that required balancing backend engineering, deployment quality, and data-driven decision support, which strengthened my ability to build maintainable systems that directly support business operations.

• Built backend microservices for the SmartInventory system using Python, Java, and RabbitMQ, helping support real-time inventory workflows across Bosch warehouse operations.

• Improved inventory retrieval efficiency by 25% and supported real-time updates across 5+ warehouse hubs, contributing to smoother operational visibility during peak usage periods.

• Developed real-time monitoring dashboards using React.js, GraphQL, and WebSockets, enabling faster visibility into system activity and issue patterns.

• Integrated LSTM autoencoders into dashboard-driven anomaly detection workflows, reducing response time by 20% and improving fault detection accuracy by 35%.

• Contributed to the Bosch Virtual Assistant chatbot by improving both intent understanding and response quality through BERT fine-tuning, BM25 retrieval, GPT-2 fine-tuning, and RLHF-informed response generation workflows.

• Helped improve chatbot classification accuracy by 30%, reduced misclassification by 25%, and automated 40% of customer queries, which improved responsiveness and reduced manual support effort.

• Gained strong hands-on experience working across backend systems, real-time data flows, dashboard engineering, and applied NLP in a production-oriented environment.