AI Engine Ecosystem
The LinkedIn Content Engine, Person Research Engine, Company Research Engine, Post-Call Follow-Up Engine, Social Omni-Post Engine and AI Readiness Review Engine.
Saizen Acuity // Since July 2025
Work Experience
From fixing the website to fixing what it was connected to.
I joined through web development and kept finding deeper problems. Pages needed analytics. Analytics exposed manual work. Manual work became automations. Eventually the job changed from making the website better to making the operation itself behave better.
01 // The Path
I started on the website. It turned out the website was connected to everything.
The internship became a full-time role in August 2025. I was still learning most of it on the job, including how a workflow can fail silently in production. More on that further down.
02 // Now
Data & AI Engineer
I build and maintain the automation layer behind Saizen's internal work and client systems: data engineering, AI orchestration, analytics and workflow architecture. The other half of the job is research, and enough business context to notice when a technically elegant answer is solving the wrong problem.
5,000+
hours saved every month across the engine portfolio
15+
automation engines, built to one standard
1,200+
tasks completed every day
+5,464%
reported LinkedIn impressions in the content system's first month
The result was not speed alone. Work became repeatable, visible, easier to scale, and able to keep producing after deployment.
Scope & Focus
What the job actually is.
-
01
Automation Architecture
Design and maintain connected systems across research, content, communication, lead generation, analytics and client delivery.
-
02
AI Workflow Engineering
Build AI agents and multi-stage workflows that collect information, reason across it, produce structured outputs and trigger what happens next.
-
03
Research & Evaluation
Find out what a problem really is, then test the answer honestly. For Bump4Life AI, that meant a 200-question benchmark against ChatGPT and Gemini, every answer graded by hand.
-
04
Data & Analytics
Connect campaign data, UTM structures, GA4, CRM records and conversion events into reporting people can actually use.
-
05
Reliability & Security
Credential protection, role-based access, audit logs, error notifications, retries and recovery, because production has opinions.
-
06
Strategic Alignment
Work directly with leadership so the technical systems serve growth, marketing, operations and client outcomes, not just each other.
Deployment Frameworks
Standard workflow templates that cut the time to deploy a new client engine from hours to minutes.
Reporting Infrastructure
Automated PDF and HTML exports for AI-generated reports, KPI summaries, recommendations and client action plans.
Campaign Intelligence
Link hygiene standards, UTM governance and campaign tracking, so attribution stays accurate across platforms.
Connected Operations
Content publishing, CRM enrichment, research, data processing and delivery, joined up with Python, JavaScript, APIs, GA4 and workflow orchestration.
Selected Projects
Some of what that looks like.
-
Applied AI & Health Data01Bump4Life AI
The AI behind Bump4Life answers specific questions about years of maternal and child health data, by retrieving the right records and calculating against them instead of feeding a model the lot.
97.0%correct on a 200-question benchmark, graded by handRead it → -
PropTech Data Infrastructure02Prestige Properties Data Feed
A fragmented property inventory turned into one structured dataset, ready for the website, Property Hive and Rightmove feed testing, without the data becoming the weak link.
457property records cleaned and standardisedRead it → -
Browser-Based Engineering03Vectis.ai PCB & Schematic Demos
Two browser demos that draw structured hardware-design data as schematics and boards you can inspect and edit, and flag any change that still needs a backend electrical check.
2browser prototypes, schematic and PCBRead it → -
Platform Engineering04Saizen Rebuild V2
The company's second-generation platform, rebuilt as a Python and Flask application rather than another set of WordPress pages, with one application serving everything from structured content.
V2Python and Flask, still in progressRead it →
03 // Try to Break It
What happens when this assumption is wrong?
A senior engineer asked me that after one of my workflows failed silently in production. The logic was clean. The world was not. Pick something that goes wrong, then run it through the kind of demo that fools you and the version that belongs in production.
What goes wrong
Which version
What happened
How it ended
What the customer receives
Validation before confidence. Fallbacks before elegance. Human review before anything irreversible.
A simplified model with an invented customer, not a live system.
04 // Where It Started
Web Developer & Programmer
My first job was rebuilding the digital layer: development, design, SEO, analytics, security, and the automations around the website. What started as improving pages became the technical base for later engines, measurement systems and client-facing services. The site stopped being a static presence and became infrastructure.
+270%
page views
+119%
sessions
+69%
visitors
+46%
average session duration
Heatmaps and GA4 replaced arguments about design with evidence. We could see where attention died, where people stopped, and what actually changed after a rebuild.
Scope & Focus
Everything a website quietly depends on.
-
01
Web Development
Built and refined WordPress and Elementor pages from the ground up, with custom CSS and JavaScript wherever the native tools fell short of the standard.
-
02
Conversion Architecture
Structured pages around clear journeys, deliberate spacing, stronger calls to action and mobile-first usability.
-
03
SEO Foundations
Metadata, schema, internal linking, image optimisation, semantic hierarchy and on-page SEO through RankMath.
-
04
Behavioural Analytics
GA4, heatmaps, scroll tracking, click tracking and event measurement, to replace design assumptions with evidence.
-
05
Security & Reliability
SSL, access control, firewall configuration, plugin governance, backups and overall platform resilience.
What I Built
A new digital foundation.
Saizen Acuity Website Rebuild
Reprogrammed the main website with a unified dark visual system, stronger responsiveness and a more deliberate brand experience.
Modular Landing Pages
Reusable frameworks for services, AI Readiness Reviews, strategic offers and future campaigns.
Dynamic News System
A structured news module that publishes live updates without compromising performance.
Automated Lead Routing
Forms connected to Slack and Google Sheets through n8n, so nobody had to carry a lead across by hand.
Analytics-Ready Architecture
A site prepared for the dashboards, automation engines, attribution systems and campaigns that came after it.
Selected Projects
Where that work went.
The internship lasted two months. The web work did not stop with it.
-
Real Estate Web Platform01Prestige Properties Website
A cleaner, more operational property platform across frontend, listings, enquiries, migration and launch, because a property website stops being “just design” very quickly.
16website amendment areas completedRead it → -
Corporate Web & Analytics02Saizen Website Rebuild
The company site rebuilt and relaunched with analytics, authority pages, lead routing and performance work, then measured properly afterwards.
-27%bounce rate after the relaunchRead it → -
Recruitment Web Platform03ByteHire Frontend
The complete client-side build for a technical recruitment brand, through repeated changes to positioning, hiring content, navigation and layouts.
~118documented revisions handledRead it →
05 // Interface to Infrastructure
The internship taught me how a company presents itself. The full-time role taught me how it moves.
Design affects data, data affects strategy, automation changes operations, and pretending those are separate disciplines usually creates worse systems.
06 // The Standard
I am much more interested in the system around the task than the task itself. If something is slow, fragile, repetitive or confusing, I want to know why, then build the mechanism that removes the problem instead of treating the symptom again.
- Not activity.
- Not technical theatre.
- Not automation for its own sake.
- Something that still works next week.