Work
Products &
projects
Real engagements — what the problem was, what was built, and what it runs on.
Featured work
KubeManta AI — Kubernetes SRE Platform
Our own self-hosted platform for running Kubernetes: real-time workload visibility, AI-driven diagnosis, and remediation that always requires human approval.
- Live workload & node visibility
- AI diagnosis — writes always need approval
- Container vulnerability scanning
- Multi-session terminal in the browser
More work
Serverless Event-Driven Applications
Migrated a legacy e-commerce platform to Azure using serverless architecture and automated CI/CD pipelines.
Security Audit and Hardening
Comprehensive penetration testing, threat modeling, and security hardening for a client's cloud environment.
DevOps Pipeline Automation
End-to-end CI/CD pipeline implementation for a BPO software company, cutting deployment time significantly.
Kubernetes Cluster Architecture
Designed and deployed production Kubernetes clusters across GKE, EKS, and AKS for enterprise clients.
Interactive demo
RAG Platform
Service Health
API & Database
FastAPI · PostgreSQL metadata
Document Storage
MinIO object store
Vector Search
Qdrant semantic similarity
AI Engine
Ollama embeddings + Groq LLM
Avg. Response
~1.2s
Storage Limit
50 MB
LLM Model
Llama 3.3
RAG-as-a-Service Demo Platform
A fully self-hosted RAG platform — document ingestion, vector search, and grounded LLM chat — built to show how AI document intelligence can be deployed for any business.
In development
Upcoming engagements
Architecture planned. Engagements forming.
In developmentEnterprise AI Chat Platform
Network-isolated enterprise chat application using Azure AI Foundry, OpenAI models, and RAG grounding — all traffic routed through private endpoints.
In developmentMLOps Pipeline & Model Serving
End-to-end ML build and inference pipeline on AWS SageMaker with automated training, versioning, and multi-environment production deployment.