Daniel Michulke
Verified Expert in Engineering
Machine Learning Developer
After finishing his PhD in artificial intelligence (general game play) at the University of Dresden, Germany, Daniel founded a company in Luxembourg for automatic timeseries forecasts. Since 2014, he's been working as an independent consultant. His projects are mostly applications of data science with the occasional development of a company's MVP. Daniel has a strong background in math and statistics.
Portfolio
Experience
Availability
Preferred Environment
Eclipse, Jupyter, Leiningen, Git, Emacs, Linux
The most amazing...
...thing I've developed is a timeseries forecasting tool that outperforms human-built models in applications such as energy load forecasting.
Work Experience
Lead Developer
Everyair GmbH, Frankfurt, Germany
- Gathered requirements from the CEO.
- Implemented the MVP back-end for charter flight search, including smart search and simple booking.
- Deployed and maintained products.
- Advised on technological risks and opportunities.
Consulting Lead Data Scientist (Remote)
Accrue, Inc., Hong Kong
- Developed the quantitative finance back-end.
- Implemented various statistics.
- Sped up calculation to make the experience smooth.
Consulting Algorithm Engineer (Remote)
Reliance JioMoney, India
- Set up a geospatial database + efficient search algorithm (using Google S2) to find shops near customers in real time.
- Set up a flexible recommender system for coupons based on Mahalanobis Distance. The system is able to explain its recommendations and can deal with missing data.
- Implemented and deployed a demo front-end on Heroku.
- Designed and implemented various security measures for payment and coupon platform. Provided mathematical proofs for payment PIN collisions.
Consulting Developer & Data Scientist
Agri-Esprit, Strasbourg, France
- Advised on infrastructure, recruitment, and feature development.
- Designed and Implemented the server side and DB.
- Developed statistical models for analysis and forecasting (e.g., what fertilizers to use).
- Aggregated and visualized business and regulatory data.
Solutions Architect (Remote)
Financial Startup (via Toptal)
- Developed a quantitative platform.
- Advised on infrastructure and recruitment.
- Connected to financial data providers and caching data.
- Generated on-demand representation of data (e.g., candle bars from tick data, spreads, back-adjusting futures, and building generic contracts).
- Built pattern mining on Stocks/Forex.
- Conducted statistical risk analysis of patterns (volatility, Sharpe-ratio, etc.).
- Built efficient calculation and representation of approx. 5 million patterns for 10k stocks and stock/commodity/FX futures.
- Served security master data via REST interface.
CTO
M & R Investment Technologies, Luxembourg
- Developed a timeseries forecasting system for energy prices (EEX Futures).
- Built automatic data retrieval (Bloomberg, CSIdata, Weather API).
- Calculated generic contracts for futures and back-adjusting.
- Preprocessed timeseries information (normalization, z-scoring, outlier detection).
- Generated automatic models (neural networks, least squares, decision trees) and automatic back-testing.
- Conducted statistical analysis of models (R-squared, Variance, Bias, MAE/RMSE, etc.).
- Built automatic optimization for self-improving models.
- Created forecast aggregation of different models (Ensemble).
- Generated portfolios via Markowitz/Black Litterman.
- Built statistical risk analysis of portfolios (Sharpe-ratio, tracking errors, maximum drawdown, etc.).
- Delivered via daily email with an Excel spreadsheet via web services.
Researcher
Neoway, Florianópolis, Brazil
- Designed and developed a framework for breaking CAPTCHAs in a fully configurable way.
- Applied several clustering, categorization, and image recognition algorithms with partially strong mathematical requirements.
- Generalized the image recognition to have a one-fits-all configurable CAPTCHA-breaking tool box.
- Developed a system for automatically classifying OCR-ed hand-written clinic records, saving approximately 6k EUR per month.
Researcher
SAP Research, Dresden, Germany
- Rewrote a prototype for database schema matching used to assist in DB migration.
- Parallelized the system and made it faster and more memory efficient by an order of magnitude.
- Led development as product owner and main developer.
- Implemented and generalized matching components that identify similarities in two different databases.
- Integrated the product in SAP NetWeaver.
Student Worker
Oehme EDV, Chemnitz, Germany
- Designed and created an MS Access Database.
- Created a full user interface.
- Created VBA macros for automatically constructing technical designs based on product numbers.
- Created a process control application in VB that communicates with an SPS system.
Experience
Matt bei Vier Gewinnt
https://www.heise.de/artikel-archiv/ct/2009/01/174_Matt-bei-Vier-gewinntNeural Networks for State Evaluation in General Game Playing
Distance Features for General Game Playing Agents
Cryptocurrency Portfolio Tracker (incl. Financial Blog)
http://www.cryptoport.netLittle Javascript, no cookies, no email, no login. This might be important in the future if the tax man wants to know from your portfolio tracker about your assets.
Professional Blog
https://dmichulke.github.ioSkills
Languages
Python, Java, Clojure, JavaScript, SQL, Prolog, Excel VBA, Visual Basic, R
Libraries/APIs
Scikit-learn, Incanter, NumPy, REST APIs, S2 Geometry Library, React, OpenCV, Node.js, jQuery
Paradigms
Data Science, REST, Database Design, Object-oriented Programming (OOP), ETL, Logic Programming, Functional Programming, Clean Code, KISS, Unit Testing
Platforms
Linux, Salesforce, Jupyter Notebook, Eclipse, JBoss, Heroku, Windows
Storage
PostgreSQL, Database Modeling, MongoDB, XML Schema, Cassandra, NoSQL
Other
Statistical Analysis, Classification Algorithms, Linear Regression, Machine Learning, Neural Networks, Fraud Prevention, Recommendation Systems, Bayesian Statistics, Mathematics, Statistics, Scientific Computing, Big Data, Bitcoin, Applied Mathematics, Multithreading, Ensemble Methods, Evolutionary Algorithms, Optimization, Image Recognition, Bayesian Inference & Modeling, Predictive Analytics, Quantitative Finance
Frameworks
ClojureScript, Hibernate, Re-frame
Tools
Microsoft Excel, Git, Emacs, Microsoft Access, Leiningen, Reagent, Jupyter
Education
PhD ("Dr. Rer. Nat.") Degree in Computer Science / Artificial Intelligence
Dresden University of Technology - Dresden, Germany
Master's Degree in Computer Science
Chemnitz University of Technology - Chemnitz, Germany
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring