Skip to main content

Josh Lennon

I studied Artificial Intelligence at the University of Manchester, I'm studying for an MSc in Artificial Intelligence at the University of Surrey, I founded GrowTrades, and I make music.

What brings you here?

Josh Lennon

Josh Lennon

About me

MSc Artificial Intelligence student at the University of Surrey, with a BSc in Artificial Intelligence from the University of Manchester and a first-class dissertation (76%). I build working AI products end to end: a solo multi-agent system that placed 28th in the Reasoning Architecture track at the Open Agent Hackathon, and a lead-generation business with paying clients. Seeking a graduate role in AI engineering, machine learning or AI consulting.

Now

  • Studying for an MSc in Artificial Intelligence at the University of Surrey.
  • Working with Mercor as an AI expert on contract.

Not what you're after? Choose again

Selected work

All work

Arbiter

Open Agent Hackathon 2026, built solo in five days: 28th in the Reasoning Architecture track.

A multi-agent system that checks a SaaS vendor's security claims against the company's own records. Agents gather cited evidence, a versioned policy makes the decision, and every step replays from a signed audit trail.

Arbiter's analyst console, running here as a demo. Pick a vendor and press Start assessment to watch the agents reach a ruling. Assessments replay runs recorded from the real engine in sample mode, on synthetic data with no models connected; the recorded live runs used Zetaris and Nemotron on 24 and 25 October 2026.

Research

  • The first page of Introduction to Deep Learning

    Introduction to Deep Learning

    Paper25 pagesNot peer-reviewed

    A mathematical paper on the theoretical foundations of neural networks, machine learning, deep learning scalability and optimisation algorithms.

    From Introduction to Deep Learning

    Cite Introduction to Deep Learning

    Plain text

    Lennon, J., & Mukherjee, A. (2025). Introduction to Deep Learning [Paper]. https://joshlennon.com/documents/introduction-to-deep-learning.pdf

    BibTeX

    @misc{lennon2025introduction,
      author       = {Lennon, J and Mukherjee, Anirbit},
      title        = {{Introduction to Deep Learning}},
      howpublished = {Paper},
      year         = {2025},
      note         = {Not peer-reviewed},
      url          = {https://joshlennon.com/documents/introduction-to-deep-learning.pdf},
    }
  • The first page of Extracting magnetic information from Kerr Microscopy images

    Extracting magnetic information from Kerr Microscopy images

    Dissertation58 pagesNot peer-reviewed

    An automated image-processing pipeline for Kerr microscopy that improves image analysis and the extraction of magnetic data, with a web application interface to support research.

    From Extracting magnetic information from Kerr Microscopy images

    Cite Extracting magnetic information from Kerr Microscopy images

    Plain text

    Lennon, J. (2025). Extracting magnetic information from Kerr Microscopy images [BSc Artificial Intelligence dissertation, University of Manchester]. https://joshlennon.com/documents/kerr-microscopy-dissertation.pdf

    BibTeX

    @thesis{lennon2025extracting,
      author      = {Lennon, J},
      title       = {{Extracting magnetic information from Kerr Microscopy images}},
      type        = {BSc Artificial Intelligence dissertation},
      institution = {University of Manchester},
      year        = {2025},
      note        = {Not peer-reviewed},
      url         = {https://joshlennon.com/documents/kerr-microscopy-dissertation.pdf},
    }
  • The first page of Iris I could identify this flower

    Iris I could identify this flower

    Report96 pagesNot peer-reviewed

    A tool demonstrating the fundamental principles of neural network computation, programmed in Visual Basic without libraries as its own object-oriented framework, and tested on the Iris dataset.

    From Neural network from scratch

    Cite Iris I could identify this flower

    Plain text

    Lennon, J. (2023). Iris I could identify this flower [Report]. https://joshlennon.com/documents/neural-network-iris-report.pdf

    BibTeX

    @misc{lennon2023iris,
      author       = {Lennon, J},
      title        = {{Iris I could identify this flower}},
      howpublished = {Report},
      year         = {2023},
      note         = {Not peer-reviewed},
      url          = {https://joshlennon.com/documents/neural-network-iris-report.pdf},
    }

All research

GrowTrades

Lead-generation websites with instant quote calculators for local trades businesses. The first client is SCB Handyman Services in Guildford, Surrey.

The SCB Handyman site, the first GrowTrades client, running here as a demo. Nothing you enter in it is sent anywhere.

Read the case study

More experience

Mercor

AI Expert [Contract], August 2026 to now

  • Evaluating and annotating AI model outputs in large-scale workflows, producing training data for model fine-tuning while keeping accuracy and throughput high alongside full-time study.

BEATLEASE

Co-founder, 2025

  • Co-founded, during the final year of my BSc, a SaaS platform where music producers connect with clients and lease them their music.
  • A lower-cost alternative for customers who felt existing platforms neglected their needs.
  • Consulted directly with clients, which I enjoyed, turning their non-technical feedback into technical work for me or another developer.
  • Full-stack development with Remix, PostgreSQL, Stripe Connect and Cloudflare R2.

About BEATLEASE

Play the footage: A figure in a white shirt walking through long grass at golden hour, beside a drift of smoke.

Music

I also make music as a hobby

I'm Too Late This TimePlays from YouTube.
Real LifePlays from YouTube.
Josh Lennon's artist profile in the Spotify app: 610 monthly listeners, with It Ain't No Big Thing as the most popular track.

Lately

All posts

Elsewhere

GitHub

Josh Lennon, 9 public repositories

1,256 contributions in the last year
GitHub contributions by month
MonthContributions
October 20250
November 20250
December 20250
January 20260
February 20260
March 20260
April 20260
May 2026208
June 2026533
July 2026423
August 20268
September 20260
October 202684
  • neetcode-submissions

    My NeetCode.io problem submissions

    Python0 stars

  • COMP34111-AI-Games-Hex-Group34

    A Monte Carlo AI agent built to compete in the game of Hex. This was a team project that scored well.

    Python0 stars

  • BEATLEASE

    A scalable platform that allows music producers to sell their beats and lease options via custom website.

    TypeScript0 stars

View profile on GitHubUpdated

LinkedIn

Josh Lennon

AI Engineer

  • Founder, GrowTrades
  • AI Expert [Contract], Mercor

View profile on LinkedIn

Get in touch

Whatever brings you here, I'd like to hear from you.