I have learnt the most as co-founder and CTO of a quantitative trading startup. Over 8 years, my role evolved from hands-on development to building and leading teams, shaping strategy, and eventually returning to in-depth technical work focused on scaling and performance optimisation.
Now, as a freelancer, I apply that expertise to projects where it’s most helpful: distributed systems, complex Python applications, optimising Python performance, cryptocurrencies, and algorithmic trading. With a strong background in mathematics, I’ve also picked up data science and machine learning – not just as a practitioner, but also as an engineer enabling both research and the productisation of ML models.
I work as a business partner to startups, scaleups, and mature companies. Talk to me if you are facing a tough scaling problem, designing new architecture, or facing a complex technological decision. I can help by analysing the problem, suggesting potential solutions, and writing prototypes or production-quality code. I provide guidance and share practical insights to accelerate your projects and break through challenges.
Can I help you with your project?
Let's chat — [email protected] & linkedin.com/in/vitamidori
Some of my work
ML research & engineering at Demokratis.ch
Demokratis transforms Switzerland’s consultation procedure on new laws (Vernehmlassungsverfahren) by bringing it into the digital age and making it more accessible. To achieve this, we need to automatically process and comprehend thousands of pages of legal text in three languages. I conduct research on the best NLP tools and state-of-the-art language models to employ, and I’m responsible for deploying these solutions into production. demokratis.ch ↗︎
Redesigning and optimising a cluster of microservices at an algorithmic trading firm
The client brought me on to rethink their architecture for distributed, near real-time processing of market data. The goal was to create a more maintainable system and resolve critical performance bottlenecks. I analyzed the existing codebase, proposed refactoring strategies, and executed a proof-of-concept refactor of the most complex microservice. Additionally, I extensively profiled their Python code and implemented several in-process performance optimizations.
Elastic compute & MLOps at an algorithmic trading firm
Highly reliant on agile, high-quality research, this company employs dozens of data scientists and researchers across multiple teams. My role was to design and initiate a company-wide framework, platform, and set of standards and best practices for data science and ML research, evolution, training, and deployment. A key aspect of this work was developing a parallel computing framework for training models on large cloud clusters.
Talks on Python software engineering
I publicly spoke about Python software engineering on several occasions, for example about static typing, real-time market visualization, or trading platform design in general. (In case you’re wondering, I used to have a different surname in the past.) I like sharing my knowledge and experience with others, and am always open to speaking at conferences and meetups.
Parallel computing & other news in Python
In addition to my conference talks, I also analyse and write about new features in Python. In 2024, new parallel execution models became possible in Python, and I built a tracker project to follow their general availability. And to summarise the main topics and the hottest GitHub projects in the Python ecosystem in the past year, I created a comprehensive 2023 summary using machine learning and manual curation. python.tips ↗︎
Consulting on trading and risk management for a crypto payment processor
The core business of this company is processing cryptocurrency payments and acting as a gateway between traditional banking and digital assets. Beneath the surface, however, lies a complex operation similar to running a market-making trading desk. I was hired to help them understand this business model and its key risks. Drawing on my experience in quantitative trading, I analysed their live trading data and assessed the efficiency of their current market operations. I delivered a report with recommendations for key improvements, as well as reproducible code for further data analysis.
