Co-Founder | Fractional CTO | Deep-Tech Builder | AI, Cloud, Security and Data Systems
I build useful technology, go deep into the engineering details, help teams make better technical decisions, and try to keep complex systems understandable enough that future-me does not complain too much.
AtharvaAI - Bartman - RefundAero - Kendr - HeroVired - YouTube
I work at the intersection of engineering leadership, product thinking, AI systems, cloud architecture, security, and developer education.
My strength is not just writing code. It is asking what should be built, why it should exist, how it should be architected, and how to help a team ship it without turning the codebase into a group therapy topic.
I enjoy the strategic side of technology, but I am still very much a hands-on engineer at heart. I like understanding the internals, trade-offs, failure modes, data flow, infra boundaries, security assumptions, and all the tiny details that usually wait patiently until production to introduce themselves.
| Area | What I bring |
|---|---|
| AI and Agentic Systems | LLM applications, RAG, multi-agent orchestration, long-running AI workflows, practical GenAI architecture |
| Cloud and DevOps | AWS, Azure, GCP, CI/CD, infrastructure thinking, reliability, deployment strategy |
| Security Engineering | Secure architecture, practical threat modeling, application security, cloud security mindset |
| Data Engineering | Pipelines, system design, ML-adjacent platforms, data-backed product workflows |
| Full-Stack Engineering | MEAN, MERN, LAMP, Django, Flask, Ruby on Rails, React Native, Flutter |
| Technical Leadership | Fractional CTO work, architecture reviews, engineering execution, mentorship, curriculum and program leadership |
Most of my current work sits around AI systems that need to operate beyond a single request-response loop, and the infrastructure, security, and reliability questions that come with them.
Areas I keep coming back to:
| Area | Questions I keep exploring |
|---|---|
| Agentic AI and multi-agent systems | How should agents plan, delegate, recover, remember, and coordinate over long-running workflows without becoming unpredictable black boxes? |
| Enterprise RAG and knowledge systems | How do retrieval quality, chunking, hybrid search, citations, evaluation, and data freshness change when RAG moves from demo to business workflow? |
| AI for DevOps and SRE | How can agents help with incident context, root-cause analysis, safe remediation, chaos experiments, and operational evidence without making production more exciting than necessary? |
| Kubernetes and cloud control planes | How do policy, guardrails, typed remediation, approval flows, GitOps, and audit trails make automation trustworthy? |
| Security for AI-native applications | How do prompt injection, data leakage, tool permissions, supply-chain risk, identity, and tenant boundaries change when LLMs become part of the execution path? |
| Cloud-native platforms | What patterns help teams balance reliability, cost, release speed, observability, and multi-cloud complexity? |
| Data systems for AI products | How should ingestion, vector stores, event streams, metadata, lineage, and quality checks work when downstream behavior depends on context? |
| Developer education and simulation | How can labs, AI tutors, sandboxes, and scenario-based learning help engineers learn by doing instead of just nodding at slides? |
| Quantum and cryptography | What does quantum computing mean for cryptography, identity, and long-term security design? |
The pattern is usually the same: read, build, break, write notes, improve the system, and pretend that was the plan all along.
I am currently building Kendr, a multi-agent orchestration platform focused on practical, industry-aligned agent workflows.
The goal is simple: make agent systems more useful for real production environments, especially where tasks are long-running, context-heavy, and operationally messy.
Like many good engineering projects, it started with a few "this should probably work better" moments and then quietly became an obsession.
Kendr is where a lot of my current thinking converges:
- agent orchestration
- reliable long-running workflows
- enterprise AI architecture
- human-in-the-loop execution
- tool use, memory, planning, and task delegation
- production patterns beyond demo-grade agents
I also spend time on open-source projects where AI, infrastructure, and operational safety meet. That is usually where things get interesting, because the happy path is rarely the whole story.
| Project | Why it matters |
|---|---|
| KubeAthrix | An open-source, pre-release Kubernetes guardrail and remediation control plane that turns security and reliability findings into reviewable, typed, approval-aware changes with verification and audit evidence. |
Not every repo here is trying to be a shiny product. Some are labs, some are teaching material, some are experiments, and some are me pulling a thread until the system starts explaining itself.
| Repository | Signal |
|---|---|
| LearningGenAI | A hands-on GenAI and Agentic AI learning path for cloud and DevOps engineers moving from "I used ChatGPT" to "I can build with this." |
| superrag | Production-minded Enterprise RAG SaaS starter with FastAPI, Next.js, PostgreSQL/pgvector, Kafka, Redis/Celery, MinIO/S3, Docker Compose, tests, and Terraform starters. |
| A2A_Demos | Runnable Agent2Agent ecosystem with planner and worker agents, task lifecycle, persistence, and a UI for visualizing interactions. |
| OpenDevOpsAgent | End-to-end A2A plus MCP distributed-agent workflow showing orchestration, tool access, resources, and structured JSON-RPC responses. |
| ChaosEngineeringAIAgent | AI SRE and chaos-engineering platform around EKS, Terraform, Argo CD, observability, safe experimentation, and agent guardrails. |
| LearningK8s | Kubernetes and EKS training labs, manifests, cluster setup, and practical cloud-native exercises. |
| LearnDevOps | DevOps learning material across deployment, Docker, CI/CD, cloud, and infrastructure workflows. |
| LearnSecurity | Cybersecurity and web-security notes with OWASP-oriented learning and lab setup. |
I try to keep one foot in strategy and one foot in the codebase, because both punish hand-waving in their own special way.
| Role | Focus |
|---|---|
| Co-Founder / Fractional CTO at AtharvaAI, Bartman, RefundAero, Kendr | Product architecture, engineering direction, AI-first systems, technical execution |
| Director of Tech Programs at HeroVired | Technology education, curriculum strategy, mentoring, industry-aligned program design |
| Public Builder and Educator | Sharing projects, experiments, and engineering lessons on YouTube |
AI Agents LLMs RAG Agentic AI Cloud Architecture AWS Azure GCP DevOps Security Data Engineering System Design Python Node.js React Django Flask Ruby on Rails MongoDB SQL Flutter React Native
Currently learning and exploring:
- Quantum Computing
- Quantum Cryptography
- Better patterns for agentic systems in production
- Security implications of AI-native applications
I am usually up for thoughtful conversations around:
- AI product architecture
- multi-agent systems
- cloud and DevOps strategy
- security-heavy engineering
- CTO advisory and technical leadership
- developer education and mentorship
- building products from zero to production
If that sounds like your kind of engineering conversation, feel free to reach out or wander through the repos. The notes are mostly organized. Mostly.