Peter Baksh Salisbury · UK
Dossier · Dr Peter Baksh

AI Engineer
& leader.

I build production LLM applications and lead AI strategy — alongside consulting, education, and talks on AI adoption. Currently running an AI platform pilot with embedded tutors and MLflow-based governance. Co-creator of podreadai.com and multiple AI tools using RAG, agentic workflows and multi-modal pipelines across OpenAI, Anthropic, Gemini and open-source models.

Name
Peter Baksh, PhD, M.Phys(Hons), PGCE
Based
Salisbury, United Kingdom
Focus
AI Leadership · Production LLM Engineering
Email
pete@petebaksh.co.uk · pdbaksh@gmail.com
Phone
+44 07717 771492
Elsewhere
LinkedIn · petebaksh.co.uk
Peter Baksh portrait
Plate 01 — Portrait Peter Baksh
Currently   AI Product and Solutions Lead, Corndel Certified   Microsoft AI Transformation Leader (AB-731) Available   for consulting, education & talks
§ 01
Summary

Building the tools —
and teaching the teams.

AI Engineer building production LLM applications and leading AI strategy — alongside consulting, education, and talks on AI adoption. Currently running an AI platform pilot with embedded tutors and MLflow-based governance. Co-creator of podreadai.com and multiple AI tools using RAG, agentic workflows (OpenAI Agents SDK), and multi-modal pipelines across OpenAI, Anthropic, Gemini, and open-source models.

Microsoft Certified Trainer and Google Cloud Generative AI Leader. Developed GenAI curricula for business leaders. PhD in computational imaging with peer-reviewed publications.

§ 02
Key Achievements

By the numbers.

Client ROI
£5m+
Delivered through Data Science projects across 80+ learners — capex transparency, infrastructure modelling, customer service optimisation.
Platform Scale
10,000+
Learner capacity on enterprise AI platform; projected £2–10M ROI over 5 years. Architect & lead AI engineer.
Training
FTSE100
Developed and delivered the Strategic AI for Leadership programme to senior executives.
Research
5
Peer-reviewed publications during PhD, contributing to novel X-ray ptychographic techniques.
§ 03
Professional Experience

From research bench
to production systems.

2025 — Present
Current

AI Product and Solutions Lead

Corndel · Remote
  • Work with stakeholders across the business to identify problems that can be solved with AI, take them through to working prototypes, and manage development of the resulting solutions.
  • Engineer and PM on enterprise AI content platform (10,000+ learner capacity), architecting Vue.js/Nuxt.js solution with embedded AI tutors, conversational assessments, and RAG-based learning tools; projected ROI of £2–10M over 5 years.
  • Implementing MLflow for AI governance: automated test case generation, bot verification, red teaming for bias detection, and instruction tuning workflows.
  • Building production AI tools using OpenAI Agents SDK: video processing pipelines, automated content generation, and multi-modal learning experiences.
  • Authored executive decision brief securing pilot funding for AI platform initiative.
2022 — 2025

Data Science and AI Lecturer

Corndel · Remote
  • Developed and delivered “Strategic AI for Leadership” programme to FTSE 100 companies, covering GenAI fundamentals, traditional AI, and change management strategies.
  • Created “Generative AI for Business Leaders” and “Generative AI for Beginners” curricula for enterprise clients.
  • Taught 80+ learners in Data Science, ML, and Python; delivered client projects generating £5M+ ROI across capex transparency dashboards, infrastructure modelling, customer service optimisation, and contract data integration.
2023 — Present
Part-time

Data Engineer

Mensch Absolute Return Fund · Remote
  • Built and maintain automated daily ETL pipeline for portfolio asset pricing and performance calculations.
  • Developed visualisation dashboards to support investment decision-making.
  • Tech stack: Python, API integrations, AWS (Lambda, Glue, Athena, ECS, ECR).
2024 — 2026
Closed

Founder

Summit-AI LTD · Remote
  • AI consultancy for SMEs seeking GenAI strategy and implementation guidance.
  • Designed and delivered 7-part “Introduction to Generative AI for IFAs” webinar series for BAT Software, reaching 10–30 financial advisers per session.
2023 — 2025
Part-time

Assessor

Academy for Digital · Remote
  • Working as an assessor for the Data Analytics 4 Standard.
2020 — 2022

Electro-optic Scientist

Defence Science and Technology Laboratory · UK
  • Developed ML prototypes (CNNs, LSTMs, Autoencoders) achieving 99%+ accuracy in controlled environments, informing strategic defence research direction.
  • Applied neural network architectures to electro-optic sensor data analysis.
2019 — 2020

Data Science Engineer

Roke · UK
  • Developed computer vision algorithms for high-resolution LIDAR data to automate jet engine measurement, improving efficiency and safety in aerospace inspection.
  • Tech stack: Python, Git, Linux.
2018 — 2019

Platform Systems Engineer

Defence Science and Technology Laboratory · UK
  • Applying machine learning techniques to image processing to investigate unique signatures within data.
  • Processing LIDAR data using point cloud and intensity analysis.
§ 04
AI Projects & Applications

Selected work.

01
AI-Native Content Generation Platform Vue.js / Nuxt.js · OpenAI Agents SDK · MLflow
Lead AI Engineer on content platform with embedded AI tutors, conversational assessments, and RAG-based learning tools. Includes MLflow governance pipeline for automated test case generation, bot verification, red teaming, and instruction tuning. Built with OpenAI Agents SDK; exports SCORM packages for LMS delivery.
02
PodReadAI Live · Gemini Flash/Pro · OpenAI
Extracts and transcribes podcasts, enabling conversational search via RAG. Built with Gemini Flash/Pro and OpenAI models. Deployed and publicly accessible.
03
Video Cropping Tool Electron · Whisper · GPT
AI-powered video editing tool using OpenAI Whisper for transcription and timestamping, GPT for intelligent break detection, and human-in-the-loop verification workflow. Automatically segments and crops video at detected breaks.
04
Explainer Video Generator Electron · GPT · Gemini · ElevenLabs · FFmpeg
Video generation pipeline: GPT for script creation from context or web search, ImageGen/Gemini for visuals, ElevenLabs for narration, timeline-based animation controls, FFmpeg for final rendering.
05
GP Diagnostic Assistant React · Cloud · Proof of Concept
Cloud-deployed clinical support tool that transcribes consultations, provides differential diagnosis suggestions, and generates follow-up correspondence. Demonstrates art of the possible for NHS GP workflows.
06
AI Marking Study Anthropic · OpenAI · Gemini · Llama / Groq
Comparative evaluation of AI marking capabilities across Anthropic (Haiku, Sonnet, Opus), OpenAI, Google Gemini, and Llama models via Groq, assessing accuracy and consistency for educational assessment.
07
LLM Fine-tuning Exploration LoRA · Gemini · Hugging Face
LoRA adapter fine-tuning for Gemini using custom datasets. Hugging Face tutorials on fine-tuning open-source models. Compared instruction tuning vs full fine-tuning trade-offs for production scenarios.
08
Enterprise Image Branding Tool Gemini · Prompt Templates
Gemini-powered image generation with company-specific prompt templates, brand guidelines integration, and persistent chat histories for consistent visual content creation.
§ 05
Skills & Education

Capabilities & credentials.

LLM & Generative AI
OpenAI API (GPT, Whisper, ImageGen, Agents SDK), Anthropic Claude, Google Gemini (Flash, Pro), Llama/Groq, LangChain, LangSmith, ElevenLabs TTS, RAG architectures, vector stores (Kendra, OpenAI, local), fine-tuning (LoRA adapters), Hugging Face, MLflow.
LLM Theory
Transformer architecture, attention mechanisms, MoE (Mixture of Experts), Grouped Query Attention, sliding window attention, tokenisation strategies.
Core Programming
Python, SQL, MATLAB, Bash/shell scripting, HTML, CSS.
AI-Augmented Development
Building production applications using LLM coding assistants (Copilot, Claude Code) with TypeScript, JavaScript, React, Vue.js/Nuxt.js, Next.js, Electron.
Cloud & Infrastructure
AWS (Bedrock, Kendra, Lambda, ECS, Glue, Athena), Azure (OpenAI Service, ML), GCP (Vertex AI), Docker, Airflow, Prefect, Git.
Classical ML
Classification, regression, clustering, neural networks (CNNs, LSTMs, Autoencoders), NLP, computer vision, Python (scikit-learn, TensorFlow).
Data & Analytics
Power BI, ETL pipelines, data visualisation, API integration, web scraping.
AI Leadership
GenAI curriculum development, AI strategy design, change management for AI adoption, responsible AI, stakeholder engagement.
Post Graduate Certificate in Education (PGCE)
2017 — 2018
Qualified Teacher Status (QTS) and graded outstanding in all teaching placements. Skills developed in people management, presentation, planning, report writing, teaching psychology and time management.
PhD in Photonics
2012 — 2016
Thesis: Ptychographic Imaging of Real Biological Samples using a High Harmonic and Synchrotron Source. Developing a novel technique for high resolution imaging, utilising the coherent properties of x-rays generated by the high harmonic generation process. The project uses computational algorithms to replace the role of the objective lens typically used in microscopy.
MPhys Physics, 1st Class Honours
2008 — 2012
Modules: Advanced Quantum Mechanics, Linear Algebra, Lasers, Classical Mechanics, Electromagnetism, Medical Physics, Atomic Physics. Masters Project: Investigation into the Photon Rise Time Dynamics of VECSEL Lasers.
Secondary Education
2001 — 2008
Ravens Wood School · UK
A-Levels (2008): Geography A, Maths A, Physics A. AS-Levels (2007): Biology A, Further Maths A. GCSEs (2006): 10 × Grade A–A*, including Maths and English Language.

Certifications

The Alan Turing Institute — Operationalising Ethics in AI (Expert)
Apr 2026
The Alan Turing Institute — Operationalising Ethics in AI (Intermediate)
Apr 2026
Anthropic — Introduction to Claude Cowork
Apr 2026
Anthropic — Claude Code in Action
Mar 2026
Anthropic — Introduction to Model Context Protocol
Mar 2026
Anthropic — Building with the Claude API
Mar 2026
§ 06
Publications

Peer-reviewed work.

§ 07
About Me · Field Notes

Beyond the professional world.

Beyond the professional world, I’m driven by a love of outdoor adventure, endurance sports, and helping others reach their potential.

i.

Teaching & Education

At heart, I’m a teacher and educator. I started my career as a Physics teacher in secondary education, where I discovered a deep passion for helping people understand complex concepts and achieve their potential. This love of teaching naturally led me into Data Science and AI training, where I now help professionals and organisations navigate the rapidly evolving world of artificial intelligence. For me, AI isn’t just technology — it’s an enabler that can help people work smarter, learn faster, and achieve more than they thought possible.

Triathlon competition
ii.

Skiing

I’m a keen skier with a background in competitive ski racing, having represented the England Ski Team and competed at national and European level. These days I’ve traded the race gates for backcountry touring, finding greater fulfilment in exploring remote mountain terrain and earning my turns. There’s something special about the solitude of a dawn skinning session followed by untouched powder descents.

Ski touring in the mountains
iii.

Running

Running is where I find my greatest sense of adventure. While I’ve competed across distances from 5K to 100km ultras, my real passion lies not in the podiums but in the journeys — the pre-dawn trail starts, the summit finishes, and the wild places that only your own two feet can take you. Whether it’s a fast road 10K or a full day in the mountains, running remains my meditation and my exploration.

Ultra running
iv.

Triathlon

I spent several years as a competitive triathlete, achieving podium finishes at both Ironman and 70.3 distances. The discipline of balancing swim, bike, and run training taught me valuable lessons about consistency, marginal gains, and pushing through discomfort — skills that translate well beyond sport.

Trail running adventure
§ 08
Athletic Achievements & Interests

The logbook.

Triathlon
1st place age group Forestman Ironman Triathlon (2013)·3rd place Ironman Weymouth 70.3 (2019)·1st place London Duathlon standard distance (2019)
Running
1st place Endurance for Life trail running series — Half Marathon Dorset (2014), Marathon Devon (2015)·1st place Continental 12 hour lightning run (2016)·1st place Bournemouth half marathon (2017), Salisbury half marathon (2019), Clarendon marathon (2023)
Skiing
England Ski Team (2010–2012)·European Dry Slope Ski Championships 3rd place (2007)·British University Indoor Ski Championships South: 1st (2011), 2nd (2012), 3rd (2010, 2014, 2015)·National University Ski Championships 1st place (2010, 2011, 2013)
Awards & Recognition. Received ‘Full Colours’, ‘Half Colours’, and ‘Club Colours’ from the University of Southampton Athletic Union. Nominated for ‘Sportsman of the year’ (2010) and ‘Special Recognition Award’ (2014) by the University of Southampton Athletic Union.
§ 09
Additional Work Experience

Before all this.

Summer Intern Gill Jennings and Every · UK
2017

Assessing current patent applications which required amendments, and altering the claims accordingly. Carrying out extensive searches of the prior art in order to correctly define the novel invention within the space of intellectual property.

Cycle Mechanic and Customer Sales The Bike Shed, Wanganui · New Zealand
2017

Building, servicing and repairing push bikes. Working on the shop floor, dealing with customer sales and transactions. Implemented automated service request emails, online forms and automated database updates.

Ski Coach Southampton Alpine Centre · UK
2010 — 2011

Coached skiing from beginners to advanced racers while completing MPhys degree. Qualified Level 1 Ski Instructor and Level 2 Alpine Performance Coach.

Let’s talk.