Daniel Podrażka · AI & Data Engineer

Building intelligent systems on robust data infrastructure.

I help asset managers, macro research firms, and trading desks build the data infrastructure and systematic processes behind better investment decisions.

Get in touch ↗ See what I've built
★ Top Rated Plus on Upwork · $80k+ verified earnings
01

About

I combine engineering depth with real markets experience, so what gets built actually matters for your P&L.

01

Market data pipelines

Ingestion, normalization, and storage for tick, bar, and fundamental data.

02

Backtesting frameworks

Rigorous checks on data integrity, execution assumptions, and lookahead bias.

03

Trading system design

Connecting research signals to execution across futures and other asset classes.

04

Technology stack evaluation

Steering firms away from overengineered solutions and vendor lock-in.

What I don't do

I'm not selling you an alpha signal or a black-box algo. I assess, design, and build the infrastructure and tooling around your strategies, so your team can research, backtest, and deploy with confidence.

02

Skills

Languages & Data

Python PostgreSQL DuckDB Vector Databases Dagster

AI & LLM

Claude API OpenAI API Deepseek Qwen PydanticAI LlamaIndex FastAPI

Market Data

Databento Massive (formerly Polygon) Bloomberg API Financial Modeling Prep Interactive Brokers (TWS API)

Infrastructure & Web

AWS GCP Docker GitHub Actions AWS Amplify React
03

Selected experience

AI-Powered Trading Infrastructure

Python · DuckDB · LLMs

Built full trading systems from scratch:

  • Integration with 10+ financial APIs for market data and intelligence
  • AI-driven fundamental and technical analysis using Claude and GPT models
  • Custom SQL indicators for advanced market analysis and signal generation
  • Semantic document search using AWS S3 Vectors
  • Custom backtesting engine with Sharpe, Sortino, and drawdown analytics
  • Low-latency real-time processing in Python and DuckDB
  • Batch pipeline orchestration with Dagster for historical analysis and reporting

AI Agent Development & Open Source

MCP · PydanticAI · Dagster

Designed and shipped agent tooling, data pipelines, and long-context systems:

  • Dagster Document Intelligence: an open-source Dagster pipeline for LLM-powered document extraction, built as a client-facing demo
  • Hierarchical Memory Middleware: a 4-level compression system for effectively unbounded AI conversations, with automatic memory management via MCP tools over history stored in DuckDB
  • Editor-MCP: a code editor built for AI-agent interaction (10+ stars, 5+ forks)
  • Custom agentic workflows with PydanticAI for structured interactions
  • Semantic parsing of long documents with LlamaIndex

Data Engineering, TrueRoll

2021–2025 · GovTech

Built robust data systems for a GovTech startup:

  • Proactive monitoring systems for fraud detection
  • Comprehensive data pipelines with built-in monitoring and analysis
  • Data-product APIs focused on integrity and real-time processing
04

Get in touch

build@daniep.com

Interested in working together, or have a system you'd like a second opinion on? Reach out.

Email ↗ LinkedIn ↗ GitHub · @danielpodrazka ↗