John Cant

Data scientist with experience in commodity and equity trading, holding a Certificate in
Quantitative Finance, and a strong software engineering background. I can design and build trading
analytics solutions end-to-end, generate systematic alpha through quantitative and fundamental
market analysis, and implement through automated trading or delivery to discretionary traders.

July 2024 – Present: Sibylla LTD, Data Scientist

- Sibylla builds machine learning equity trading systems based on fundamentals. They boast
excellent and consistent YoY returns, and may be about to partner with a major fund.
- Contributed a number of improvements to the trading strategy
- Improved explainability, leading to further improvements in strategy performance
- Continually learned about the equity space, and proposed a steady stream of further
improvements
- Achieved a 10x improvement in runtime in a critical place

Oct 2019 – March 2024: Shell, Data Science Contractor

Project 1:

- Originally hired to assist with the development of a bunkering monitoring tool to help estimate low sulphur vs high sulphur fuel oil demand during the IMO2020 transition, which helped traders make $1b during that time [1]

- Quantified and improved detection accuracy to the point where we could nowcast the demand accurately enough to be used after the transition, which was not previously thought possible.

- Automated all manual steps, automated deployment, added CI, environment separation etc.

- Expanded to the rest of the world

Project 2:

- Developed another advanced software/data science product to detect specific types of unusual vessel behaviour with a price impact. This was responsible for generating $20m of business value in a single quarter.

Project 3:

- Developed quantitative trading signals for crude freight trading (Physical and TD3C FFAs)

- Communicated data science results with the aid of a streamlit dashboard, which also made them reproducible and easily extendable in future.

- Demonstrated modelling accuracy via metrics

- Devised a trading strategy and a trading simulation, in order to demonstrate profitability, under replicated market conditions (low liquidity, bid-ask spread)

- Our product has indirectly helped traders make £30m in 2023.

General / Misc:

- Worked with traders and analysts to understand their modelling and data requirements

- Technologies used: Python, SQL, Azure, Cython

- Data includes maritime data, economic data, alternative data, structured, unstructured, high quality, low quality, largest dataset was 1.6bn rows, ~16TB

- Saved £46k in consulting fees by avoiding an unnecessary cloud migration

- Five times recipient of Shell Five Star honours.

Jan 2019 – July 2019: Photobox, Data Science Contractor

6 month contract to maintain and enhance a recommender system. Improved total run time from 12h to 35m, added extra inputs, introduced industry-standard software engineering best practices such as testing, CI, separation of environments, containerized deployment, and developer documentation.

May 2016 – Nov 2018: VEEPIO, Data Science Contractor

During this time I designed and built a scalable real-time analytics pipeline for a mobile analytics product now used by three NFL teams. I also analysed this data to provide insights, automating scaling and extending this over time. I selected video tracking algorithms for our use case, and ported a state-of-the-art video tracking algorithm to iOS using an early release of CoreML.

June 2015 – May 2016: Sententia Ex Data, Consulting Data Scientist

During this period I was a freelance software and data science consultant. Due to my wide ranging experience, I have been able to win clients in London, Cologne, New York, and Abu Dhabi. The work has been increasingly varied and interesting, with one project demonstrated at COP21, and another used by a major fashion brand. Since May 2016, my services have been engaged repeatedly by VEEPIO.
 

June 2015 – Jan 2023: AirPublic, Director

AirPublic was a mobile and static sensor network that measured air quality in cities. I co-founded AirPublic as part of two separate teams that joined forces after coming first and second in two hackathons. I took on responsibility for its hardware and software. AirPublic had previously obtained grant funding and was accepted onto Bethnal Green Ventures accelerator programme. It went through been through four hardware iterations, and its clients included schools and councils.
 

Mar 2015 – May 2015: Zoaring Adaptive X, Research and Development officer

Provided the numerical and computational part of a psychopysiological research skunkworks. Worked closely with a cognitive psychologist to devise experiments, collect data, test hypotheses, analyse results, and guide further research. Experiments required computer vision and signal processing.
 

May 2014 – Mar 2015: AMEE Ltd., Data Scientist and Software Engineer

Devised a statistical model to estimate emissions for companies in the UK, responded to ad-hoc requests for data, informed visualizations, imported data from disparate sources, contributed to website development, and delivered an important project for the Connected Digital Economy Catapult.
 

Apr 2013 – Apr 2014: Boppl Ltd. and Juper Ltd., Software Engineer

Worked in a dynamic team of engineers, building the backend for Boppl, an industry leading mobile ordering app. Built a payments API and admin API, assumed responsibility for devops, assisted with PCI compliance, advocated TDD/BDD, and introduced Git. The app was rolled out in venues across London.
 

Feb 2013 – Apr 2013: Reevoo Ltd., Software Engineer

Participated in discussions around refactoring the data model.
 

July 2011 – January 2013: BeHiring Ltd., CTO

BeHiring was a software startup which successfully obtained funding and support from BetaFoundry accelerator (Oxford) and Oxygen Accelerator (Birmingham) in 2011. BeHiring was nominated OnRec Newcomer of the Year for 2012. I cowrote our software using Ruby on Rails, participated in investor and client meetings, and pitched our business to the CEO of PerTemps, the largest company in the UK recruitment industry. During this time I learned Ruby on Rails and kept up to date with cutting-edge web technology. I also managed web designers, graphic designers, and interns, conducted usability tests, and responded to feedback from our existing users. We released our minimum viable product and sought further investment.
 

September 2007- May 2011

University of Oxford. I studied Engineering Science, specialising in Information Engineering, a broad topic that includes computer vision, machine learning, control theory, and cognitive systems. My fourth year project was to build a desktop application for automatic photo organisation and summarization using PHOW (Pyramid Histogram of Words), a state-of-the-art algorithm prior to the deep learning revolution.

September 2007- May 2011, MEng in Engineering Science, University of Oxford, 1st.

I studied Engineering Science, specialising in Information Engineering, a broad topic that includes computer vision, machine learning, control theory, and cognitive systems. My fourth year project was to build a desktop application for automatic photo organisation and summarization using PHOW (Pyramid Histogram of Words), a state-of-the-art algorithm prior to the deep learning revolution.