Lead ML and DataOps Engineerporidhi.io
July 2024 - Present• Architected cloud-native EKS/Kubernetes infrastructure for AI agent deployments with auto-scaling for optimized resource utilization• Built distributed model serving architecture for efficient cross-cluster inferencing with SLO/SLI tracking for model performance and reliability• Served as Engineering Lead and Product Engineer across 3 products: SentinelOps (agentic SRE tool with eBPF), PUKU Editor (AI-powered VS Code fork), PUKU CLI (AI coding agent in terminal) — collectively serving 1,000+ active users• Launched Poridhi AI Studio in beta — a cloud-native GPU-accelerated development platform providing on-demand workspaces for AI model training and serving• Owned TensorCode as Product Owner, a browser-based CUDA/Triton/PyTorch IDE with NVIDIA A100/T4 GPU access designed to teach GPU programming from fundamentals to kernel-level optimization• Developing automated monitoring and alerting systems for ML infrastructure• Implementing RLHF-based fine-tuning workflows to study alignment trade-offs between reward modeling quality and serving latency in production inference pipelines
Junior Machine Learning EngineerJanata Wifi (janatawifi.com)
Jan 2023 - Dec 2023• Built Karigor AI Copilot agentic workflow — 98% reduction in troubleshooting time for network engineers and support staff, resolving incidents in minutes vs. hours• Built Anomaly Detection System for CXM Dashboard using TensorFlow, ELK Stack, and Azure ML Pipelines — 50% improvement in root cause analysis efficiency• Improved RAG system performance 30–40% via LLM caching and OSS model orchestration with LangChain — 35% cloud cost reduction• Architected hierarchical AI agent monitoring system (LangChain, LangGraph) and migrated entire ML workflow to Azure