Agentic AI Portfolio

I take agentic AI products from prototype to production and run the LLMOps behind them, from deployment and tracing to guardrails and evals.

AI Helpdesk Agentic Flow

An agent that grants access by policy, and can show why.

The gateway and dashboard every model call goes through.

Each skill once, with the projects that use it
Skill	AI Helpdesk Agentic Flow	LLMOps Console
Applied AI
Agentic AI with LangGraph	Yes	No
MCP tool server	Yes	No
RAG pipeline	Yes	No
Qdrant vector search	Yes	No
Versioned prompts	Yes	No
LLM guardrails	Yes	No
Conversation evals	Yes	No
LLM platform
LLM gateway with model routing and fallback	No	Yes
LLM-as-judge evals with a CI eval gate	No	Yes
PII redaction	No	Yes
Semantic caching	No	Yes
Token cost tracking and budgets	No	Yes
LLM APIs
Anthropic Claude API	No	Yes
OpenAI API	No	Yes
Azure OpenAI	No	Yes
Backend
Python (async)	Yes	Yes
FastAPI	Yes	Yes
SQLAlchemy	Yes	No
Alembic	Yes	No
PostgreSQL	Yes	Yes
Redis	Yes	Yes
Qdrant	No	Yes
Security
RBAC	Yes	No
JWT auth	Yes	No
Multi-tenant isolation	Yes	No
Audit log	Yes	No
Observability
OpenTelemetry	No	Yes
Langfuse	No	Yes
Prometheus	No	Yes
Grafana	No	Yes
Frontend
Next.js	Yes	Yes
React	Yes	Yes
TypeScript	Yes	Yes
Tailwind	Yes	No
Cloud and DevOps
Docker	Yes	Yes
GitHub Actions	Yes	Yes
Kubernetes (Kustomize)	Yes	Yes
Terraform	No	Yes
Testing
pytest	Yes	Yes
Playwright	Yes	No
Trivy image scans	No	Yes
k6 load tests	No	Yes
vitest	No	Yes