fred's site.
thoughts
books

Trading experience summary for EasyA

Hello. My name is Fredrik, but you can call me Fred.

Phil asked me to send over my specific experience in the quant domain.

I was among the first people to start at TickUp in Sweden. It’s an AI prop shop owned by a very wealthy person. Basically, he wanted to use AI to invest for him. We grew it from 3 to 30 people ish. I worked there May 2020 to November 2021. I left to take some time off (since TickUp did quite well), and then tried to do a startup (it didn’t do so well). Now I work as a freelance AI contractor with a gig at Rillion, but saw your post in my feed and couldn’t resist seeing what happens.

Can’t disclose PnL during my tenure, but what I can say is the owner had to publicly declare his income in the Swedish tax registry and came out as #1 in all of Sweden. Owner would be Claes here (Swedish article): https://omni.se/borshaj-fran-arboga-drog-in-305-miljoner-pa-ett-ar/a/lVLKnk

Our org structure was 3 teams:

  • Quant team (research)
  • Quant dev team (building a trading platform that quants could use)
  • Engineering team (ingesting data, and building systems for that)

I worked cross-disciplinary in all 3 teams. I was the guy who could speak both mathematics with the quant PhDs, and engineering with the system engineers. Being able to be a bridge in “both domains”, my primary focus was mostly spent on the quant dev team building the trading system. I was also responsible for implementing strategies using NLP, as my Masters is CS focused on AI and natural language processing. It was somewhat rare to employ strategies in the pre-LLM era; I suspect they must work much better now, I’m super excited to explore that further.

At TickUp we went all the way from filling 7 colo lockers with machines (leveraging about 6PB of various market data and alt data), to building a trading system leveraging the data, to research successful applied AI alphas trading on the markets.

TickUp experience

Software Engineering

  • Co‑author of TickUps proprietary algorithmic trading system that traded on global financial markets in mainly Python and some parts in C++.
  • Wrote and integrated an event‑sourced backtesting engine with useful quantitative finance utilities and performance metrics.
  • Authored the internal risk metrics library and service.
  • Technical skills: Python, Kafka, event sourcing, Numpy, Kubernetes, Docker, Grafana, gRPC, REST, C++, Parquet, Arrow

Data Scientist (Quant)

  • Quantitative/algorithmic trading strategy research on large amounts of data in Python.
  • Exploratory research and evaluation of alternative data sets.
  • Seasonality research on alternative data sets.
  • Feature and predictive power analysis of alternative data sets.
  • Trained deep neural networks to faster estimate option volatility surfaces.
  • Trained deep neural networks to emulate complex (costly) price estimation functions faster than numerical reference implementations.
  • Authored NLP trading strategies based on various news and social media data.
  • Technical skills: Python, NLP, statistics/math, Pandas/Numpy/Jupyter, TensorFlow, Prophet, neural networks, Linux, Docker

Data Engineering

  • Helped build and manage the bare metal infrastructure co‑lo (7 full server lockers with hundreds of various machines/clusters and switches). I’m good at using tiny screwdrivers now.
  • Responsible for the high‑performance ML research environment, data needs and data availability for all quantitative researchers.
  • Ingested many live and historical data sources from different data vendors for use in our live trading systems and backtesting simulations.
  • ETL jobs with several data sets
  • Advanced partitioning and indexing of underlying data for massive query speedups
  • Collaborated on building a 6 petabyte CEPH cluster for historical exchange tick data and how to partition it to be used in our live trading and strategy research environments.
  • Technical skills: Big data/data lakes, ETL, partitioning/indexing, Postgres, Linux, bash, Ansible, Kubernetes, Docker, Parquet, Arrow, avro, CEPH, co‑lo hosting, networks

Hobby capital:

Software Engineer & Data Scientist

  • Authored a proprietary trading system (live trading and backtesting) in Python for automated short‑term equity trading on Nordic exchanges.
  • Managing a data lake with various alternative data and intra‑day exchange data.
  • Wrote automated scrapers for various Swedish news outlets.
  • Created novel NLP‑based intra‑day trading strategies using Swedish news, machine learning and intra‑day exchange data.
  • Technical skills: Python, event sourcing, statistics, nltk, spaCy, Parquet, Numpy, Docker, LLMs, AWS, Dremio, monitoring