Custom tools that
actually work
I design and build production software tailored to your business — custom internal tools, workflow automation, and AI integration. No templates, no throwaway prototypes. Systems your teams actually adopt.
A clear process,
predictable results
No surprises. Here's exactly how we go from problem to production.
Solutions for your
real problems
Whether you need AI-powered automation or a custom business tool — I build what your team actually needs.
Real systems,
real results
Built, deployed, and running in production today — anonymized, but every metric is real.
Editorial Intelligence Platform
A publishing team needed structured research content across hundreds of entities in multiple languages. Manual process: 4+ hours per article. Now: under 45 minutes with higher consistency.
- Replaced a 4-hour manual research process with automated, source-cited intelligence gathering
- Quantified data completeness for the first time — team sees exactly what's missing across 500+ entities
- Full audit trail on every data point — critical for editorial credibility
- Platform keeps working even when AI services go down — no single point of failure
Multi-Module Business Platform
A fast-growing brand needed rock-solid tools for their entire operation — financial modeling, design collaboration, media management, and AI-assisted content. 6 integrated modules, one platform, running in production.
- Custom financial calculator — pricing, margins, break-even, multi-channel sales modeling
- AI design studio — extracts design DNA from references, generates blueprints, renders photorealistic variations with natural language refinement
- AI-powered knowledge base with writing assistance — cut content production time by 60%+
- Full observability across all modules — usage analytics, cost tracking, performance monitoring
Before / After
Editorial Intelligence
Producing structured research content for a 500-entity knowledge base.
- Manual research across dozens of sources
- Copy-paste into documents, inconsistent formatting
- No way to measure data completeness
- Translation outsourced, weeks of delay
- Automated research pulls from a synced knowledge base — no more tab-switching across dozens of sources
- AI-assisted drafting with built-in writing, editing, and translation — one interface, not five
- Automated completeness scoring tells you exactly what's missing before you publish
- Same-pipeline translation — minutes instead of weeks of outsourced delay
AI Design Studio
From weeks of external designer back-and-forth to same-day AI-assisted iteration.
- External designers with multi-day feedback loops
- Expensive per-revision pricing — teams hesitate to iterate
- Vision gets lost in translation between briefs and deliverables
- No connection to brand guidelines or existing design assets
- AI extracts design DNA from reference images and generates production-ready variations
- Natural language refinement — "make the dial darker" — no design skills needed
- Every output grounded in your brand guidelines via connected knowledge base
- Unlimited iterations at fixed cost — experiment freely
From concept to production
Real tools I've built — dashboards, AI studios, automated pipelines. Every interface designed for the people who actually use it.
Real-time KPIs your team checks every morning
Senior engineer, not an agency.
15+ years of software engineering, from startups to enterprise. I don't sell shiny demos — I deliver systems that run in production, with security, monitoring, documentation, and audit trails.
My approach: understand your business first, choose the technology second. AI isn't an end in itself — it's a lever when well-integrated into your existing workflows and tools.
Recent work: a 6-module business platform spanning financial modeling, design collaboration, media management, and AI-powered editorial workflows — all running in production. Whether the solution needs AI or not, I ship tools that teams rely on daily.
Common questions
Custom business tools, AI-powered workflow automation, internal platforms, and data pipelines. If your team has a process that doesn't scale or software that doesn't fit, I build the solution from scratch — designed around how you actually work.
No. AI is a powerful lever when it fits, but many of the tools I build don't use AI at all — financial calculators, collaboration platforms, operational dashboards. I choose the technology that solves the problem, not the other way around.
Four phases: Discovery (audit processes, define KPIs), Prototype (working POC validated with your team), Deployment (integration, security, monitoring), and Optimisation (metrics tracking, adjustments, training). You get a free initial analysis before any commitment.
It means the software ships with security hardening, monitoring, documentation, error handling, and audit trails. No throwaway prototypes — systems your teams can rely on daily, with resilience built in from day one.
Primarily TypeScript, Next.js, React, Node.js, and Python for AI/ML workflows. Databases include PostgreSQL, SQLite, and Redis. But the stack is always chosen to fit the problem — not the other way around.
Have a process that
needs fixing?
Tell me about the problem you're trying to solve. I'll get back within 24h with a free initial analysis — no commitment, no sales pitch.