- Operating as a commercial project manager, directly managing a team of 52 contractors within complex multi-site operations.
- Managing contracts, purchase orders, cost reports, forecasts and invoicing — driving accuracy in financial tracking and compliance with commercial agreements.
- Building advanced trackers with automations and advanced formulas, giving the team a clear, live picture of current statuses and forecasts.
- Advanced BidCraft into specialised AI products: policies & procedures bots, tender-analysis bots and tender-summary bots, with further bespoke solutions in development.
- Exploring wider AI implementations across the Bovis business; recognised for efficient delivery of project milestones and managing competing priorities through an operational transition.
About Me
I'm an AI engineer and automation specialist — I find the inefficiencies hidden inside workflows and companies, then build the AI that removes them.
At Lendlease I spotted a slow, repetitive problem in how the team answered tenders, so I built an AI tool trained on our library of historic bid answers. It delivers evidence-based responses with supporting case studies and a feedback loop for continuous improvement — producing more competitive bids and saving around a week's work for three to four people on every bid it supports. At Bovis, that tool became BidCraft and grew into specialised products for policies & procedures, tender analysis and tender summaries.
Day to day, I operate as a commercial project manager at Bovis: 52 contractors, contracts, purchase orders, cost reports, forecasts and invoicing — the discipline of actually shipping. Alongside that, I build advanced trackers with automations and advanced formulas that give teams a clear, live picture of current statuses and forecasts. My foundation is engineering: Apache Spark and Scala pipelines, Hive and MySQL for big data, EMR deployments, and full-stack builds with Python and React.
I'm passionate about AI and how it can make the workforce more efficient — better outputs, easier days, for everyone. I'm at my best where engineering meets delivery: spotting the repetitive, manual work nobody questions, automating it away, and making everyone's day a little easier. I'm looking to grow further into AI development and automation, bringing both the technical depth and the delivery discipline with me.
Experience
- Designed and built an AI-powered bid-response tool trained on historic bid answers, delivering evidence-based responses complete with case studies — the origin of BidCraft.
- Implemented Power Automate workflows and built multiple Power BI dashboards, improving project reporting and analysis.
- Created project systems from scratch: snag-list tracking, an asset tracker, a DSR log, a centralised project-experience database and a client feedback system.
- Played a key role in RISQE audit preparation — evidence collection, documentation validation and quality compliance. Certified in NEC4 Foundation, CITB Core Trades Awareness, Asbestos Awareness and IIF.
- Built an end-to-end Spark (Scala) pipeline over continuous real-world API data — nested-JSON flattening, regex cleaning, broadcast-optimised joins and robust null handling.
- Delivered MySQL-to-HDFS ETL with Sqoop incremental loads, landing in partitioned, Avro-backed Hive tables queried with optimised HiveQL.
- Shipped readable, maintainable code throughout — well-labelled stages, clear commentary and Avro handling via the Databricks Spark-Avro library.
- Team-built a niche e-commerce platform for tailors to showcase and sell cultural clothing, with profiles, login and reviews.
- Pioneered automated image tagging with the Google Cloud Vision API; frontend in React & Next.js, backend on Firebase & Firestore.
Earlier (2018–2022): retail at JD Sports (store record for most Crep Protects sold in a day) and STEM tutoring — at AE Tuition and running my own practice, Jay's Tuition. The communication and people skills from those years are what I use daily managing 52 contractors now.
Flagship — BidCraft
Case study AI in a live business, not a demo
An AI bid-writing tool, trained on years of historic bid answers, that became a product family.
The problem
Answering tenders meant repeatedly rewriting responses to similar questions, hunting through past projects for supporting evidence. Slow, inconsistent, and it pulled senior people away from strategy.
What I built
An AI tool trained on our extensive library of historic bid answers. Ask it a bid question and it returns a high-quality, evidence-based response with supporting information and relevant case studies from past projects — with a feedback system for continuous improvement.
The outcome
On average, around a week's work saved for a team of three to four people on every bid it supports — roughly 220–450 hours returned to the business every month. It's also used to critically evaluate draft answers, sharpening what's written — and it has helped win several bids. At Bovis it evolved into specialised products: policies & procedures bots, tender-analysis and tender-summary bots.
Featured build Finding inefficiency beyond AI
SharePoint Redesign & Information Architecture
Saw a SharePoint estate that had grown clunky — tangled folders and hard-to-find documents — and led a complete redesign. Ran workshops with the real users to understand how they actually work, rebuilt the structure to be intuitive and better organised, and delivered a cleaner, nicer-looking UI. The result: things are found faster, and the whole team works more efficiently.
Selected builds
High-Availability Compute Cluster
A Linux cluster engineered for continuous service through redundancy and failover — Nginx web serving and load balancing, SSH remote administration, UFW firewalls, and Prometheus + Grafana monitoring. Resilience validated across failure scenarios.
Spark Data Pipeline
Production-style Spark (Scala) application refining continuous real-world API data: nested JSON flattening, regex cleaning, broadcast-optimised joins, null strategies and temporal enrichment.
Hadoop ETL Platform
End-to-end integration from MySQL into HDFS via Sqoop with incremental loads, landing in partitioned, Avro-backed Hive tables queried with optimised HiveQL.
Study Planner + Blog Platform
A Django web app that generates personalised study plans from student syllabuses — secure login, profile management, tailored revision/practice/memorisation formats, a daily completion log, and an interactive blog for collaboration.
Cultural E-Commerce Hub
Google Hackathon build: a marketplace for tailors selling cultural clothing, with automated image tagging via Google Cloud Vision for smarter search, plus full user profiles and reviews.
Arduino Embedded Systems
Led software and design across a series of Arduino builds: a gesture-controlled table centrepiece with hidden compartments and LED ambience, a working traffic-light system model, and a camera-based reader that plays tunes from detected component patterns.
Capabilities
AI & Automation
Big Data
Cloud & Infrastructure
Programming
Web & Product
Data & ML
Delivery
Credentials
Education
2021 — 2024
BEng Computer Systems Engineering
Royal Holloway, University of London · High 2:1. Project manager for the winning team in termly projects three times.
2016 — 2021
Herschel Grammar School
A Levels including Computer Science (A*) and Mathematics (A).
Certifications
- APM Associate
- Databricks Academy — Generative AI Fundamentals
- Get Started with Databricks for Data Engineering
- Databricks — Fundamentals of the Lakehouse
- NEC4 Foundation
- CITB Core Trades Awareness
- Asbestos Awareness
- Incident & Injury Free (IIF)
Client feedback
Why hear it from me — when you can hear it from the people I've delivered for?
Jay's communication and support has been nothing short of exemplary… he has demonstrated a calmness that others could learn from. Jay is a great person to work collaboratively with — his behaviours are something Bovis should be proud of. He conducts himself and represents the organisation in the highest order.— ██████ · Finance, Network Rail
Since you joined, I've found everything to be running smoothly and efficiently… your responses to my queries have been consistently prompt and helpful, and I really appreciate the timely processing of my timesheets — it makes a big difference.— Contractor · Network Rail framework
The information I was supplied with by Jay and the process to get the contract up and running was clear, precise and straightforward… Jay proactively engaged with the payment company and kept us fully informed. The contract seems to be running smoothly and well managed thanks to the help from Jay.— Contractor · Network Rail framework
You are extremely professional, supportive and informative… your support has been invaluable. Without your support I do not believe issues would get resolved or taken as seriously as they do. I appreciate everything you do and am very happy with the support you provide.— ██████ · Contractor
Jay is a committed and collaborative young professional who works closely with others to achieve strong outcomes. He communicates regularly and effectively at different levels of the organisation, and consistently builds positive relationships across the team.— Client feedback · NRT framework
One of Jay's key strengths is his reliability and willingness to engage… his positive attitude and collaborative style contribute to a productive working environment — supportive, flexible and outcome focused. A strong contributor with a clear commitment to success.— Client feedback · NRT framework
08 Get in touch
Let's build something.
Whether it's an AI role, an automation challenge, or a workflow full of inefficiencies that needs someone who can both build and deliver — I'd like to hear about it.
kodalijayadev@gmail.com