MechAI← Home

Curriculum vitae

Adam KrysztopaPhD

Senior Data Scientist · AI/ML Team Lead

Warsaw, Poland

Senior Data Scientist and AI/ML Team Leader with a PhD in Applied Physics and a prior career in aerospace engineering. I combine engineering intuition with production-grade AI to solve predictive maintenance, time-series forecasting, digital-twin, and retrieval problems in industrial environments — from end-to-end ML delivery and model interpretability to systems design and turning expert domain knowledge into scalable software.

Currently AI/ML Team Lead at STX Next and running MechAI consulting alongside — open to advisory and consulting work on industrial AI.

~13 yrs
Total experience
~5 yrs
Data science
~2 yrs
AI/ML leadership
~8 yrs
Aerospace eng.

Experience

STX Next

Apr 2024 – Present

Senior Data Scientist / AI/ML Team Leader

  • Lead AI/ML workstreams across predictive maintenance, digital twins, document AI, RAG, and agentic AI; act as technical lead and mentor in an internal ML reskilling program.
  • Design AI/ML solution architectures — data flows, model workflows, validation layers, observability, deployment — behind maintainable Python interfaces.
  • Build Remaining Useful Life (RUL) estimation for Oil & Gas (regression + classification) with a focus on interpretability and Azure ML delivery.
  • Built a production hybrid search engine (BM25 + PGVector dense retrieval + cross-encoder reranking) and a multimodal RAG system (shared image+text embedding space, visual QA via vision LLMs).
  • Built decision-support models for energy commodity traders — a 3-day rolling RTM-vs-DAM allocation model for the ERCOT market (Sharpe-scored) and a month-ahead gas price-index selection model.
  • Apply LLMOps & evaluation practices — Arize Phoenix, NeMo Guardrails, golden-dataset evaluation pipelines with regression checks.

MechAI

Apr 2024 – Present

Founder · Consulting

  • Independent data-science consulting focused on predictive maintenance, lifecycle optimization, and engineering analytics for thermal and mechanical systems.

Schneider Electric

Jul 2021 – Mar 2024

Senior Data Scientist / Senior Data Analyst

  • Built sales & orders demand forecasting (classical, ML, probabilistic) for the global finance function — 12-month rolling forecasts blending internal drivers and macroeconomic indicators, with stakeholder-interpretable output a hard requirement.
  • Developed the SWDA/SWADA seasonal-adjustment methodology — deseasonalization and working-days correction preserving MoM/YoY growth for financial planning.
  • Built data-engineering and validation pipelines with Python, Databricks, PySpark, S3, and Prefect.

GE Aviation

Apr 2013 – Jun 2021

Thermal Systems Design — Senior / Advanced Lead / Lead Engineer

  • Thermal systems, heat transfer, and secondary-flow design across commercial and defense-adjacent aviation programs, including the GE Catalyst turboprop.
  • Led technical scope in high-stakes, regulated environments — cross-functional execution, structured documentation, certification-oriented validation.
  • Built a strong foundation in safety margins, mission-critical reliability, and structured engineering problem-solving.

Skills

Languages

  • Python
  • SQL
  • asyncio (advanced)
  • TypeScript (basic)

ML / DL

  • PyTorch
  • TensorFlow
  • scikit-learn
  • XGBoost
  • LightGBM
  • HuggingFace
  • Nixtla
  • Pandas / NumPy
  • Computer vision (U-Net)
  • A/B & champion–challenger

GenAI & agents

  • PydanticAI
  • LangGraph
  • LlamaIndex
  • LiteLLM
  • AutoGen
  • vLLM
  • Ollama
  • MCP servers & clients
  • Multimodal Document AI
  • Multimodal RAG (VQA)

Search & retrieval

  • Hybrid (BM25 + dense)
  • Cross-encoder reranking
  • RAPTOR
  • Table-RAG
  • Multimodal retrieval

Cloud & MLOps

  • Azure ML (~1.5y)
  • Databricks (~3y)
  • AWS (Bedrock, Lambda, S3)
  • Docker
  • GitHub Actions / GitLab CI

LLMOps & observability

  • Arize Phoenix
  • NeMo Guardrails
  • Langfuse
  • RAGAS / ARES eval
  • OpenTelemetry
  • structlog + Grafana

Backend

  • FastAPI
  • SQLAlchemy
  • PostgreSQL / PGVector
  • Redis
  • APScheduler
  • Event-driven async

Domains

  • Predictive maintenance & RUL
  • Digital twins
  • Time-series & probabilistic forecasting
  • Forecastability triage
  • Advanced RAG
  • Energy trading / commodities
  • Demand forecasting
  • Aerospace & defense

Education & certification

  • PhD, Applied Physics

    Warsaw University of Technology

    Degree

  • MSc, Applied Physics

    Warsaw University of Technology

    Degree

  • Postgraduate, Data Science

    Białystok University of Technology

    Degree

  • Architecture & Systems Engineering

    MIT xPRO

    Certification

Writing & speaking

Writing

  • Production AI & RAG vs fine-tuning
  • Predictive analytics & predictive maintenance
  • Oil & Gas and aviation use cases
  • Industrial AI implementation guides

Speaking

Data Science Summit · Engineering meetups

Predictive maintenance · Digital twins · Elastic RAG

This is the condensed version. For specifics, references, or a summary tailored to a particular role or problem — let’s talk.

Start a conversation→