AI engineering · Wrocław, Poland

AI systems,
engineered for the real world.

HPC-AI Lab combines research rigor with hands-on software engineering to design, optimize, and scale AI systems that work beyond the demo.

Led by Piotr Miłkowski
Scientist · Software developer · AI engineer

ENGINEERING / END TO END SYSTEM ONLINE
01 Workload Understand
02 AI system Engineer
03 Infrastructure Scale
RESEARCHRigor
SOFTWARECraft
DELIVERYImpact
15+Years in software
development
PhDComputer Science
Dissertation under review
A*Research presented at
CORE A* conferences
ML/NLPDeep technical
specialization

01 / Expertise

Where AI ambition
meets engineering reality.

Focused support for the technical challenges between a promising model and a dependable production system.

01PERFORMANCE

Inference & optimization

Make AI workloads more efficient through systematic profiling, model-serving improvements, and infrastructure-aware optimization.

  • Performance and cost analysis
  • Inference architecture
  • Workload profiling
02INFRASTRUCTURE

AI & GPU infrastructure

Design practical foundations for demanding AI systems across cloud and on-premise environments, with scalability built in.

  • Infrastructure architecture
  • Cloud and on-premise platforms
  • Scalability planning
03PRODUCTION AI

LLM & data systems

Turn language models and organizational data into useful, maintainable products grounded in sound software architecture.

  • RAG and GraphRAG systems
  • AI agents and workflows
  • Intelligent document processing
04LEADERSHIP

Technical leadership

Bridge research, engineering, and product goals while helping development teams build stronger AI capabilities.

  • Hands-on technical guidance
  • AI research and development
  • Developer upskilling

Beyond the core

Additional capabilities

  • 01AI strategy & roadmapping
  • 02MLOps & observability reviews
  • 03Technical due diligence
  • 04Research partnerships
  • 05Workshops & mentoring
  • 06AI security & governance
Platforms & stack
  • Azure
  • Azure OpenAI
  • GCP / Vertex AI
  • AWS
  • NVIDIA DGX

02 / Selected work

Selected systems.
Research in practice.

AI platforms, applied research, and language technology built for financial services, data-intensive products, and scientific collaboration.

01FINANCIAL SERVICES · DOC AI

Intelligent Document Processing at BNY

Led the design and development of an AI platform for automated data extraction in a regulated financial environment, combining document understanding with production monitoring and MLOps.

Focus
Architecture, product ownership, AI platform leadership
Stack
Azure, Azure OpenAI, OCR, NLP, GenAI, microservices
02BIG DATA · DEEP LEARNING

StockBrief

Co-developed a system that analyses financial indicators and investor sentiment with big-data algorithms, generating stock-market forecasts with deep neural networks.

Focus
Model design, sentiment pipelines, evaluation
Stack
Python, deep learning, time-series modelling, opinion mining
03SUBJECTIVE NLP · RESEARCH

Personalized NLP

Researched how individual users perceive emotions, humour, and hate speech, developing architectures that combine text encoders with user-specific Human Bias representations to model subjective judgements.

Focus
Model architecture, experimentation, publications
Stack
Transformers, user embeddings, multi-task learning
04LANGUAGE INFRASTRUCTURE · SEARCH

AZON, CLARIN & QTravel

Contributed to national and European projects building language resources, search systems, and intelligent travel search based on natural-language understanding.

Focus
NLP pipelines, information retrieval, language resources
Stack
Information retrieval, knowledge bases, recommendation logic

03 / Approach

Less theatre.
More engineering.

A clear path from understanding the problem to delivering a system your team can own and evolve.

  1. 01

    Understand

    Start with the workload, constraints, and actual business objective—not a predetermined tool.

    DISCOVERY
  2. 02

    Design

    Evaluate the trade-offs and shape an architecture that fits the system, team, and operating environment.

    ARCHITECTURE
  3. 03

    Engineer

    Build, measure, and refine hands-on, with production requirements part of the work from day one.

    DELIVERY
  4. 04

    Transfer

    Document decisions and share knowledge so the solution remains maintainable after delivery.

    OWNERSHIP

Flexible collaboration

Ways to work together

04 / About

Scientist’s curiosity.
Engineer’s discipline.

I’m Piotr Miłkowski, a data scientist and software developer working at the intersection of artificial intelligence, natural language processing, and high-performance systems.

My experience spans academic research, software architecture, and technical leadership. I enjoy difficult problems where rigorous thinking and practical engineering have to meet.

Connect on LinkedIn

SELECTED EXPERIENCE

Callstack

Senior AI System Engineer

PwC

Lead AI Engineer

BNY

AI/ML Leader · Application Development Manager

Wrocław University of Science and Technology

Machine Learning Researcher · Language Technology Group G4.19

EDUCATION

PhD in Computer Science
Dissertation under review · Wrocław University of Science and Technology

05 / Contact

Have a difficult
AI problem? Let’s talk.

Tell me what you’re building, where it’s getting stuck, and what a good outcome looks like.

HPC-AI Lab · Wrocław, Poland