Engineering leader working across platform infrastructure and applied AI.

I have spent fourteen years at IBM India Systems and Development Lab, building hyper-converged infrastructure, software-defined storage and the Kubernetes platform underneath them, along with the teams that ship it.

Today, I lead various areas covering Virtualization, Multicluster and AI solutioning.

I am also an Executive PhD candidate in Operations Management at IIM Udaipur, researching artificial intelligence and technology management.

• lazy-senior-dev: the senior engineers your AI agent can be. Three persona plugins that review the change an agent is about to write and refuse the write until it is right: grumpy-reviewer reviews the diff, paranoid-sre reviews the deploy, and tenured reviews it against the repository's own history. One ruleset renders to fourteen agents, an MCP server and a GitHub Action, and every number comes from a benchmark you can rerun. The cast
• ocm-mcp-server: personal side project. Puts policy checks, approval and audit between an AI agent and a Kubernetes fleet through an Open Cluster Management hub, so the agent never holds a kubeconfig. Docs
• ibm-fusion-mcp-server: maintainer of a fork of containers/kubernetes-mcp-server, extended with IBM Fusion tools.
• ai-roadmap-365: nine-course open-source AI curriculum, from computing fundamentals through designing, evaluating, securing and deploying production AI systems. 154 of 365 lessons published, in progress. Read the course
• blogs: long-form writing on IBM Fusion, Kubernetes, MCP and agentic operations. Read them
• About: personal site. One hand-written HTML file, no framework, no build step, no tracker.

• Platform engineering: Kubernetes, multicluster and fleet operations, hyper-converged infrastructure
• Storage and data resilience: software-defined storage, converged storage appliances, backup and recovery at enterprise scale
• Applied and agentic AI: judgment about where an agent earns its place, how it reaches production systems under least privilege, approval before writes and a full audit trail
• Observability: led the architecture of a platform observability framework, built in Go on Prometheus, Loki and Elasticsearch
• Identity and authentication: Active Directory, LDAP, NIS, AD with RFC2307, OpenStack Keystone
• Engineering leadership: building squads from the first hire through technical direction to first production customer deployment

• Six granted US patents in storage and systems
• Two IBM Redbooks: IBM Spectrum Scale Security (REDP-5426) and unified file and object (REDP-5113)
• One IEEE paper, 2012
• Talks at OpenStack Summit Tokyo 2015, IBM Storage Strategy Days and IBM user-group sessions

• Site: sandeepbazar.github.io
• LinkedIn: in/sandeepbazar, the best place to reach me
• YouTube: Tech Horizon Hub

Outside work I run long. Longest single-stage event so far: 300 km.