Bristol, UK · Open to grad / new-grad ML & DS roles

Hi, I'm Josh.
Data Scientist & Applied ML Engineer.

I build practical ML systems for messy real-world data — from high-recall sepsis classifiers to RAG agents and transformers from scratch. I care a lot about interpretability and shipping simple solutions to complex problems.

AI Engineer Intern · Prentice Wealth ManagementMSc Data Science · University of Bristol

✉ Email ↗ GitHub ↗ LinkedIn ↗ Kaggle

01Experience

What I'm working on right now

AI Engineer Intern · Prentice Wealth Management

Current

Applying ML and LLM tooling to wealth-management workflows — building practical AI features end-to-end from data to deployment.

LLMsPythonRAGProduction AI

02Featured work

Three projects I'd love to talk through

APPLIED ML — HEALTHCARE

Sepsis Onset Prediction

High-recall ML pipeline to flag early sepsis onset in ICU patients. Tuned for the cost-asymmetry of false negatives, with SHAP-based feature attribution to keep predictions inspectable for clinicians.

LightGBMSHAPImbalanced dataHealthcare

DISSERTATION — INTERPRETABLE AI

XHAIL — Symbolic Reasoning System↗

Extended an Inductive Logic Programming framework (eXtended Hybrid Abductive Inductive Learning) in Python + Clingo. Fixed prior bugs, added new functionality, and explored how rule-based learning improves explainability over black-box models.

PythonClingoILPInterpretability

FROM-SCRATCH — DEEP LEARNING

Decoder-Only Transformer, From Scratch↗

Built a GPT-style decoder-only language model from first principles in PyTorch — attention, training loop, tokenisation — to actually understand the internals rather than just import them.

PyTorchTransformersNLP

03More projects

A wider sample

Twin AI↗

Retrieval-Augmented Generation chatbot built on Llama + LangChain. Document chunking, embedding store, conversation memory.

RAGLangChainLlamaOllama

Marine Conservation App↗

Mobile app for divers to ID marine species and log sightings. Built for the Bristol University Diving Society — won 1st in cohort for software engineering.

FlutterDjangoAWS1st place

AI Agents — Scotland Yard↗

Adversarial game-playing agent (minimax + heuristics) and pathfinding implementations for the Scotland Yard board game.

PythonSearchMinimax

Game of Life on AWS↗

Distributed-systems coursework — parallelised Conway's Game of Life across AWS instances and benchmarked across configurations.

GoAWSDistributed

Kaggle↗

Notebooks and competition submissions — practising clean EDA, feature engineering, and disciplined model selection.

NotebooksEDA

Local Weaver Gallery↗

Production Next.js site for a traditional weaver — designed, built, and deployed end-to-end.

Next.jsFrontendLive site

04Toolkit

Languages, libraries, and tools

Languages

PythonJavaSQLHaskellJavaScript / ReactGo

ML / AI

PyTorchTensorFlowLightGBMscikit-learnLangChain (RAG)LLMs / Transformers

Stats / Data

PandasNumPyMatplotlibSeabornSHAPR

Tools / Infra

GitDockerAWSJupyterNext.jsLinux

05Education

Academic journey

MSc Data Science · University of Bristol

Ongoing — 2026

Statistical Computing & Empirical Methods · Large-Scale Data Engineering · Visual Analytics · Intro to AI & Text Analytics · Data Science Methods & Practice

BSc Computer Science · University of Bristol

2:1 — 2025

Machine Learning · Computer Vision · Applied Data Science · Algorithms · Imperative & OOP · Software Tools · Computer Architecture

A Levels · Oakham School

2022

Mathematics A* · Further Mathematics A · Physics A · Computer Science A

06Get in touch

Let's talk

I'm actively looking for full-time data science and applied ML roles starting after my MSc finishes. Happy to chat about any of the work above — or about anything tangential and interesting.

✉ jeverett.post@gmail.com ↗ LinkedIn ↗ GitHub

Hi, I'm Josh.Data Scientist & Applied ML Engineer.