Rishabh Agarwal

• LangGraph Production Agents: Built stateful agent architecture using LangGraph with Docker Compose, PostgreSQL 16 + Redis 6 for production-grade streaming, pub-sub, and conversation persistence. Designed scalable LLM response patterns and state management. Deployed across 3 production workloads handling multi-turn reasoning with LangChain OpenAI integration. Tech Stack: Python | LangGraph | Docker Compose | PostgreSQL | Redis | LangChain OpenAI • Google Distributed Cloud Integration (GCP-Led): Implemented GitOps-based multi-cluster Kubernetes deployment orchestration (RootSync/RepoSync) enabling 50+ edge locations with Config Sync validation. Integrated Honeywell ecosystem with GCP Anthos bare metal infrastructure via Cloud Interconnect + Shared VPC to bypass corporate proxies. Reduced deployment errors by 99%. Tech Stack: Kubernetes | Anthos | GitOps | Config Sync | GCP Cloud Interconnect | GitLab CI | nomos • OpenClaw Memory Systems: Worked with persistent knowledge graph architecture via Graphiti MCP server + Mem0 memory framework with long-context LLM agent capabilities. Built OTEL instrumentation for observability and tracing. Showed 10x context retention improvement. Tech Stack: Python | Graphiti | Mem0 | LangChain | Docker | OTEL | MCP Protocol • BMS Multi-Phase LLM Pipeline: Optimized cost-driven pattern discovery using Azure OpenAI GPT-4.1 with iterative sampling & coverage estimation logic. Engineered checkpoint system for 100-hour+ jobs to ensure fault tolerance with Pandas data pipelines. Consolidated 600+ patterns to 50 actionable rules, cutting API costs by 65%. Tech Stack: Python | Azure OpenAI | Pandas | Pydantic | Checkpoint-based ETL • Spark RTM Embedding Pipeline: Reduced end-to-end latency in Databricks Spark Real-Time Mode architecture via Event Hub → PDF parsing (unstructured library) → embeddings. Tech Stack: Databricks Spark RTM | Azure Event Hub | Azure Storage Blob | Python | AI Embeddings

• Architected a RAG-powered chatbot using Python, MCP (3 servers with 8 tools & a LangGraph-based Client) to process 3 peta-bytes of data stored as Delta Live Tables in Azure Databricks, automating 20+ workflows for interactive analytics on industrial datasets. • Deployed microservices to Azure Kubernetes via GitHub Actions and Octopus, implemented CI/CD fixes across 11 Agile sprints and collaborated with 9+ interns; mounting MCP tools on FastAPI endpoints and exposing RESTful services to power dashboard features. • Developed Spark run-based MCP tools for SQL generation (98% accuracy), SQL validation, matplotlib & Plotly-based charting, and Cron job scheduling; reducing manual SQL code, data checks, and charting by 83%, enabling on-demand analytics from user prompts.

• Engineered storage-aware energy optimization system for ML workloads by implementing 4ms granular disk activity monitoring, analyzing ResNet-50 training which consumed 210W during peak GPU utilization, resulting in precise measurement of storage power states and 12.5% reduction in storage-related carbon emissions. • Developed custom Python monitoring infrastructure to track disk activity patterns in ML training pipelines, measuring energy variations between SSDs (0.06W idle, 3.5W active) and HDDs (2-3W idle, 4-5W active) while evaluating carbon impact using MLPerf benchmarking suite. • Investigated storage optimization strategies for ML inference workloads by analyzing I/O patterns, demonstrating SSDs have 2.91 kg CO2e/TB embodied carbon compared to HDDs' 1.2 kg CO2e/TB, while enabling 35x faster data access for real-time inference operations.

• Conducted 80+ hours of one-on-one help sessions, resolving technical issues related to OpenGL and WebGL implementation • Organized 5 "Programming with TA" sessions, with an average attendance of 75 students, boosting project submission quality by 22% • Graded 900+ programming assignments and quizzes within a 48-hour turnaround time, maintaining a 90% accuracy rate in feedback

• Directed technical initiatives as the President of Google Developer Student Clubs, enhancing peer collaboration. • Organized 12 engaging workshops on emerging technologies, including TensorFlow and Flutter, with an average attendance of 75 students. • Fostered a culture of learning and teamwork, empowering students to explore innovative tech solutions.

Implemented planning of coding boot camps for underserved educational institutions to bridge social disparities in terms of money and gender.

•Developed a machine learning model that accurately predicts crop yield trends for 5,000+ local farmers, enabling them to adapt their agricultural practices to help combat climate change while maintaining profitability while improving crop health trends dashboard. • Designed a Python-based data pipeline using scikit-learn and custom feature selection, processing ~16 TB of agricultural data daily. • Implemented privacy-preserving techniques to handle sensitive data, ensuring GDPR compliance for 5,000+ data points in midwest.

• Participated in the NSF I-Corps Site UPRM accelerator program to validate a software-based innovation. • Employed the Lean LaunchPad methodology to systematically test key assumptions about technology and customer needs. • Collaborated with mentors and industry experts to refine business strategies and enhance product-market fit.

• Developed a React-based dashboard for real-time energy monitoring, helping 5 client companies save an average of $120,000 annually. • Optimized SQL queries, significantly reducing database response time from 1.2 seconds to 300 milliseconds. • Integrated REST and GraphQL APIs to enhance app functionality with live equipment health monitoring and failure forecasting.

• Actively represented the interests of international students at the University of Minnesota, advocating for their needs and concerns. • Collaborated with the Office of Student Affairs and International Students Scholar Services to develop initiatives enhancing student success. • Facilitated awareness campaigns addressing key themes impacting international students, resulting in increased engagement and support.

• Developed and implemented marketing strategies to enhance the popularity of campus dining options. • Assisted in hosting promotional events that increased student engagement and awareness of dining services. • Improved communication efficiency between dining services and students, leading to a more informed campus community.