I build practical Python/FastAPI backend systems around AI governance: agent workflows, scoped tool access, policy simulation, approvals, audit logs, redaction, incidents, observability, regression testing, and operator-facing control layers.
My experience combines real-world internal business engineering through DANIELOZA.AI, production-shaped GenAI/backend projects, and AI automation work across internal workflow domains. The common thread is backend systems that are reviewable, auditable, secure, and useful in practical business operations.
- AI governance platforms with scoped access, policy simulation, approvals, incidents, replay, traces, and operator workflows
- RAG backends with ingestion, route transparency, citations, and evaluation surfaces
- Security and policy layers around agent/tool execution
- FastAPI backends with SQL, Docker, testing, and production-minded structure
- Operator-facing internal tools instead of one-shot AI demos
| Period | Focus |
|---|---|
| Aug 2021 - Sep 2022 | IT operations, reporting, customer-facing support, ticketing workflows, SQL basics, and internal technical processes. |
| Oct 2022 - Oct 2023 | Enterprise IT automation, chatbot-related solutions, SQL/API-style workflows, documentation, troubleshooting, and team-based technical delivery. |
| Nov 2023 - Feb 2024 | AI automation prototypes and business workflow systems, including Brand Insight Engine and Automation Control Plane. |
| Mar 2024 - Mar 2025 | Team-based AI backend and internal agent-style workflow systems: controlled tool access, approvals, audit logs, observability, and human-in-the-loop review. |
| Apr 2025 - Present | DANIELOZA.AI backend and AI automation work for small and mid-sized businesses around Wroclaw, including Danex operations tooling, reporting, invoice workflows, Telegram operations, exports, backups, and AI/OCR-assisted processing. |
Internal Salon Operations Platform
Through DANIELOZA.AI, I build and maintain backend and automation systems for Danex and other small to mid-sized local service businesses, used in daily operations. The systems support appointment workflows, invoice handling, daily revenue tracking, reporting, Telegram-based operations, CSV/PDF exports, backup workflows, and practical AI/OCR-assisted document handling.
This is real-world engineering with real users, operational needs, direct feedback, maintenance responsibility, and business constraints.
Start here if you are reviewing my GitHub for backend, AI governance, RAG, and agent-safety work.
| Rank | Project | Why it matters |
|---|---|---|
| 1 | Regulated AI Agent Platform | Main flagship project. A regulated AI assistant platform with source-bound RAG, scoped tools, policy checks, approvals, audit logs, prompt-injection resistance, Redis rate limits, Docker, Kubernetes, and security tests. |
| 2 | AGIP - Agentic Governance Intelligence Platform | Broader governance platform for autonomous agents: scoped access, policy simulation, approvals, redaction, incidents, regression testing, observability, and enterprise governance layers. |
| 3 | MCP Security Gateway | Specialist security gateway for MCP/tool execution with deterministic policy checks, approvals, rate limits, redacted audit logs, and incident creation. |
| 4 | Agent Control Plane | Operator-facing control plane for AI runs with approvals, incidents, replay compare, trace graphs, exports, and runtime reliability notes. |
| 5 | Danex RAG Service | Focused hybrid RAG API with ingestion, route transparency, citations, SQL-backed answers, query history, and evaluation-oriented design. |
| 6 | AI Workflow Observatory | Observability companion for AI-assisted engineering workflows with sessions, risk signals, verification quality, and cost visibility. |
| Project | What it demonstrates |
|---|---|
| Agent Governance Gateway | Focused gateway with registration, human approval, short-lived scoped tokens, revocation, multi-tenant isolation, PII redaction, audit logs, policy checks, and a premium dashboard. |
| Automation Control Plane | Backend-first FastAPI control layer for automation workflows with tenant auth, approval queues, usage limits, audit logs, and operator review surfaces. |
| Agent Runtime Control Tower | Runtime control tower with policy enforcement, approval routing, incidents, PostgreSQL-backed history, and Redis-backed runtime state. |
| Agent Regression Lab | Evaluation backend for agent regressions with scenario registries, run diffs, and safe replay previews. |
| Inference Readiness Advisor | CLI for planning local LLM inference across runtimes, quantizations, and workload scenarios. |
Python backend engineering
FastAPI and REST APIs
SQL, SQLite, PostgreSQL concepts
RAG pipelines and ingestion
Agent governance and control surfaces
AI runtime safety and observability
Approvals, audit logs, governance
Operator-facing internal tools
Python FastAPI SQLAlchemy SQLite Docker
Pytest RAG FAISS LangChain REST APIs MCP
TypeScript React Tailwind CSS
I am strongest in systems that need clear control surfaces: ingestion, retrieval, approval flows, runtime visibility, auditability, failure handling, and backend logic that stays understandable as the product grows.
I optimize for software that is useful in production-adjacent settings, easy to reason about, and structured well enough to evolve without collapsing into chaos.