Data scientist · Prague
Daniel Večeřa
I do data science and alternative music festivals. Mostly the first: I work on recommendations for Oneplay at TV Nova and build analytics agents that answer business questions with evidence you can check.
danielvecera.com · linkedin.com/in/1vecera · github.com/1vecera · x.com/1vecera
Selected work
-
October 2026
Starwatch
Overall champion and Social Media Deep Research topic winner out of 38 projects at Agents 0.0.7 “From Dusk Till Dawn”, Hackathon #01, Prague, 8–9 October 2026.
Maps what candidates in the 2026 Czech municipal elections in the ten largest cities posted on social media and how much attention it got, now alongside the official results. Built solo overnight, with every post linked back to its source.
-
2026
Mluva
Dictation for Omarchy Linux · open source
Speak a rough idea, then polish or restructure it with a rewrite model while every original word stays available. Speech and rewriting engines are chosen separately: ElevenLabs Scribe, a local speech model or a compatible API.
-
2026 – now
Agentic Analytics
Analytics agents for management questions
Agents that answer the next management question with evidence you can inspect: shared business definitions, a governed path to the data, and answers that keep what was measured apart from what is only plausible. Built on the client’s own stack and owned by their team, starting with one measurable use case.
Also a community meetup in Prague, in English, on Thursday 5 November 2026, which I organise with Josef Vodička and Ivan Vodička.
-
2022 – now
RichView
Co-founder
A presentation layer for agents. The agent writes Markdown with validated YAML blocks for charts and tables; a Rust engine renders it as a typeset HTML report with real SVG charts. Coming soon and open source at release; the playground already runs in the browser.
The first RichView, which I co-founded in 2022, went through Antler’s early-stage programme in Oslo.
Open source
More on GitHub.
-
Mouse & Touchpad for Omarchy
2026
An Omarchy bar plugin for mouse and touchpad speed, with automatic saving, backups and undo.
-
evoctl
2026
A command-line tool and three MCP tools that give scripts and AI agents a direct line to Evolution API, an open-source WhatsApp API.
-
daniel-vecera-ai-tools
2026
Write an agent skill once and install it to Claude Code, Cursor, GitHub Copilot, ZCode and Hermes from one source.
-
Voice Scribe
2026
Real-time voice-to-text for VS Code with live rewriting, on ElevenLabs Scribe or Google Cloud Speech-to-Text.
-
barinbar
2025
A small Python package for bar-in-bar and treemap bar charts in Matplotlib, published on PyPI.
Experience
-
2025 – now
Senior Data Scientist
TV Nova · Prague
Recommendations for Oneplay, TV Nova’s streaming service: part of the team that built its transformer-based recommender system. Led colleagues to take a churn-prediction model from idea to production. Taught AI-assisted coding workshops for colleagues.
-
2026 – now
Agentic Analytics
Independent · Prague
Developing analytics agents that answer management questions with evidence you can inspect. Co-organiser of the Agentic Analytics meetup in Prague.
-
2024
Data Analyst, Revenue Management
Groupon · Prague
Analysed marketplace performance and what drives it with BigQuery SQL, Tableau and Google Sheets.
-
2023 – 2024
Data Scientist (contract)
Top-20 pharmaceutical company · Germany, remote
Worked with McKinsey consultants on agricultural software: time-series prediction in Python, connecting the data-science modules to the backend, writing tests and helping the team ship the app.
-
2022 – 2023
Co-founder & CEO
RichView · Prague and Oslo
An AI platform that made business content and interactive charts easier to create. Selected for Antler’s early-stage programme in Oslo. Led a team of three, ran strategy and business, wrote the Python SDK and a GPT-4 chart backend, delivered a paid proof of concept and held 50 enterprise interviews.
-
2019 – 2022
Data Analyst, Front Office
VARO Energy · Switzerland
Analytics for the oil trading desk: more than 20 Power BI dashboards and pricing workbooks. Led two data scientists building a Prophet price forecast that cut local pricing-forecast error by more than 20%. Member of the energy-transition strategy project with company leadership and BCG. Organised an internal Python user group. Joined after a 2019 internship.
-
2018 – 2019
Intern, AI Solutions
PwC · Berlin
Showed that machine learning could work for electricity trading with a neural-network reinforcement-learning prototype.
Education
-
2017 – 2019
MSc Business Administration
Humboldt-Universität zu Berlin · grade 1.3 (1.0 is best)
Econometrics, statistics and data science. Thesis: interpretable machine learning linking US-listed companies’ financial statements to their future earnings. Visiting student at the University of St. Gallen in 2019.
-
2014 – 2017
BSc Economics and Business Administration
Technische Universität Dresden · grade 1.8
Accounting, controlling and statistics, taught in German. Double degree in Economics and Management with the University of Trento, taught in Italian.
Skills
- Modelling
- Recommender systems, churn prediction, time-series forecasting; PyTorch, scikit-learn
- Data
- Python, PySpark, SQL, Databricks, BigQuery, dbt and MetricFlow
- AI engineering
- Agents and MCP servers; AI-assisted development with Claude Code, Codex and Cursor
- Analytics
- Power BI, Tableau, Excel; commodity markets and trading analytics
- Certificates
- Professional Scrum Master I and Professional Scrum Product Owner I (Scrum.org, 2020)
- Languages
- Czech (native), English (C2), German (C2), Italian (B1)
Beyond work
- Co-organised the workshop area at Waking Life, a music festival in Portugal, 2024–2025.
- Enlargement manager at Junior Enterprises Europe, helping two new student consultancies start, 2018–2019.
- German Debating Champion, foreign-speaker category, 2015.
- Yoga, contemporary dance and electronic music.
Contact
The quickest way to reach me is a direct message on LinkedIn or X.