Oxford Nanopore Technologies

Aug 11th, 2023

ONT is a global leading biotechnology company that develops and sells nanopore sequencing devices.

In light of their vision to further enable the accessibility of sequencing devices to ‘everyone, everywhere’, I worked with them to develop multiple AI and ML projects across their business, namely; A ML-driven MQL system, an NBA Recommendation engine and the world’s fastest and most accurate browser-based Basecaller.

Background & Objectives

Across Browser Basecaller, Market Qualified Lead and Next-Best-Action Recommendation Engine



ONT provides DNA sequencing to customers using a sequencing device and a basecaller application. However, there are several usability issues and high technical requirements to setup and use them, leading to a need for additional infrastructure and customer support. Providing this on consumer devices via a browser basecaller was the desired solution.


MQL System

ONT was using a SaaS product for machine learning in their digital marketing, but it did not meet their expectations due to lack of flexibility and explainability. I worked with them to develop a custom AI tool with reusable APIs for lead scoring integration and dashboard functionality for effective tracking of KPIs.


NBA Recommendation Engine

ONT were using an unsatisfactory third-party software to personalise content on their websites, which did not cater to their specific industry needs. I worked with them to build a solution that would provide personalised user journeys and tailored recommendations to improve customer engagement.


Basecaller Objectives

  • Develop a browser-based basecaller application that uses AI machine learning browser APIs to improve usability for customers.
  • Reduce technical proficiency requirements for customers to set up the new basecaller application.
  • Ensure data security and privacy for customers by enabling sensitive data retention in the new basecaller application.


MQL Objectives

  • Develop a flexible AI tool for lead discovery, categorisation, and prediction with a focus on high-value leads.
  • Implement reusable APIs for seamless integration of lead scoring across various client systems, including Salesforce.
  • Include dashboard functionality for effective tracking and visualisation of key performance indicators (KPIs) over time.


NBA Engine Objectives

  • Develop a personalised content recommendation model tailored to the biosciences industry.
  • Enable sequencing of articles and recommendations based on user data and buying journey.
  • Automate data processing and integration with ONT’s CRM system for efficient content management.



Architect front-end, platform and device layer integrations and communication.

Machine Learning

Strategic direction of the custom ML model development and productionisation.

ML model porting from server-side to client-side. ML solution integration and governance to dictate the best approach to achieving business outcomes.

Frontend Development

Manage and support the development team with the creation of the web frontends based on the high-fidelity designs.

Backend Development

Platform and system-level design decisions including the strict data requirements and multi-tier API schema development.

Stakeholder Management

Work with stakeholders across the management team to agree on multiple project requirements, build the cross-platform team and deliver the projects into production.

Cloud Integrations

Integrating cloud-based telemetry, support and more.

Cloud services agreement for testing and productionisation of the solution



We developed and deployed a productionised browser-based DNA sequencing application that uses their existing server-side models ported to the browser.

– The solution achieved 1.1 million DNA sequencing samples per second on a consumer-grade MacBook Pro.

– Machine Learning models achieved between 97% – 99.8% accuracy depending on the model chosen by the user.

The solution has significantly improved customer experience and data privacy whilst simultaneously reducing costs.


The AI lead generation solution helped ONT’s marketing and sales teams to focus on more reliable sources of new business, resulting in higher quality leads with a greater propensity to buy.

The inclusion of revenue as a feature in the retrained machine learning models resulted in a 91% predictive accuracy rate for ‘likely to buy’ and a 95% predictive accuracy rate for ‘customer fit’, enabling the marketing team to confidently grade leads and resulting in 4x more marketing qualified leads turning into sales qualified leads. The sales team has also gained significant efficiencies by focusing their outreach efforts on high-quality leads with a good customer fit and high likelihood-to-buy.

NBA Recommendation Engine

Delivered a new user recommendation engine that offers personalised content recommendations to website visitors based on their interests and buying readiness. This has helped them establish their website as a reliable source of information on biosciences.

As a result, ONT saw a 200% increase in CTR per article, leading to a higher number of views and more time spent on their website, and a consistent pipeline of new potential customers.

Tools & Technologies

This project branched into many varying aspects of engineering and technologies, and a few tools to aid in greater productivity.

  • HTML5
  • CSS3 (SASS / LESS)
  • JS (Vanilla / OO)
  • React
  • NodeJS
  • Responsive
  • Mobile
  • UX / UI
  • Architecture
  • GitHub
  • SSH
  • SVN / GIT
  • Rust
  • WASM
  • Bash
  • Python
  • AI
  • Data
  • AWS
  • Azure
  • Tensorflow
  • ONNX
  • VS Code
  • Microsoft Office Suite