19 years designing for the people in the backstage of enterprise services — depot workers, finance teams under audit pressure, lab technicians, children in clinical assessments. I move between IC craft and design leadership depending on what the problem needs. The work shows up in business outcomes: fewer errors, lower costs, services that don't need training manuals.
Not from lack of talent — the teams I watched were smart. Nobody was mapping the full journey. Nobody was aligning the backstage to the front stage. Nobody was translating the gap between what the business thought the service was and what the person using it actually experienced. That gap is still what gets me out of bed nineteen years later.
From 2012 I ran a design consultancy practice through Rappit.io — not a single employer, but thirteen years of enterprise client engagements across Europe and the Middle East. Valeo, Syensqo, Newport Tank, eConnect, Omoda, Hunkemöller, Sanquin: each a separate client, a separate sector, a separate set of constraints. 80,000 daily users across Fintech, Supply Chain, and Healthcare. €5M+ in contract renewals. I hold PMP Elite, ESG Practitioner, and Google UX certifications alongside an MBA in Information Systems and a B.Sc. in Computer Technology. The credentials are fine. The 85 stakeholder interviews and 40 hours shadowing warehouse workers in steel-capped boots are what actually informed the work.
Across those engagements I've led distributed design teams spanning 4 countries and 6 time zones — 12 designers in total, across Syensqo's post-spin-off transformation (6 designers, Netherlands/Belgium/France/Morocco), Rappit's product design practice (1→4, grown and retained), and cross-functional design coordination across Valeo's 8 global R&D centres. The DesignOps frameworks I built weren't for teams of four. They were built to scale.
Based in Coimbatore — async-first since before it had a name. I run 9am CET workshops from IST and have been delivering for clients across Europe and the Middle East for over a decade. I show up having done the research, argue for the user in C-suite rooms, and measure the work in outcomes you can verify.
Make the AI's reasoning visible — not its confidence score.
Users don't reject AI because it makes mistakes. They reject it because they can't see where it's certain. The eConnect heatmap tripled adoption at lower accuracy than a competing design.
Decide where humans belong before the AI is built, not after it fails.
Correction flows and override affordances designed in from the start raised eConnect engagement from 34% to 89%.
Every touchpoint a user has is part of the trust architecture.
The heatmap worked. The account manager layer didn't — nobody briefed them on what an accuracy dip meant.
Confidence indicators and override flows belong in the spec from day one.
In regulated environments, surfacing uncertainty honestly isn't compliance overhead — it's the condition under which the product gets used at all.
One mid-level designer on my team was technically sharp but struggled to communicate decisions to stakeholders. Instead of prescribing a process, I started bringing her into C-suite syncs as my note-taker — then gradually shifted her to presenter. Within six months she was running design reviews independently. She's now a senior lead at a product-led company in Amsterdam.
The lesson: exposure beats coaching. Give people the room, then get out of the way.
Six-stage methodology. Three leadership commitments. One signature capability — human-in-the-loop AI design. Each backed by an outcome I can name.
Most AI product failures aren't model problems — they're trust and interpretability problems. I design the layer between the AI output and the human decision: confidence interfaces, exception flows, and the handoff moments that determine whether people trust the system or override it.
I applied DMAIC to Rappit.io's design process before it was called DesignOps. The result: half the rework in a single quarter, designers spending time on hard problems instead of revision loops. PMP Elite and Lean Six Sigma aren't credentials I list — they're how I think about design at scale.
I built career ladders and mentorship structures before posting a single job. One mid-level designer who struggled to present to executives is now a senior lead in Amsterdam — because I moved her into C-suite reviews as note-taker, then presenter, then owner. That's the model: exposure over coaching.
I've designed for children who can't tell you what's wrong, depot workers in safety gloves, and lab technicians under ISO audit pressure. Accessibility is never an afterthought in that context — it's the constraint that makes the design real. Screen-reader task completion from 62% to 99.2% on a national healthcare portal, through WCAG CI/CD gates baked into the process.
When I started in 2006, I thought good design meant a clean interface. I measured success in pixel precision and user satisfaction scores. I was proud of my work and most of it was wrong — not technically, but strategically. I was solving the visible problem and ignoring the system behind it.
Around 2014, after a project at a logistics firm where a beautifully designed interface failed because the backstage process it relied on was broken, something shifted. The depot workers loved the screen. The service still didn't work. That's when I understood: most design problems aren't visual problems. They're organisational problems wearing a UI costume.
Now I spend as much time on what doesn't show up in Figma — the handoff between actors, the undocumented process, the warehouse workaround that actually works — as I do designing the interfaces themselves. A beautifully designed product still fails if the backstage it relies on is broken. That's why I map both: the service that surrounds a product, and the product that delivers the service.
Usability research, UI development, BFSI compliance design. Measuring success by how clean it looked and whether users could complete tasks. Getting good at the craft of the screen.
The logistics project that broke me. Started blueprinting frontstage and backstage together. Realised the interface was often the least interesting part of the problem — the process behind it was where everything actually went wrong.
Facilitating the room with 45 business unit heads who all believed their service was most critical. Understanding that alignment is a design problem. The blueprint isn't the output — the conversation it forces is the output.
AI is changing what service design can see — pattern recognition at scale, hidden dependencies in complex workflows. But the human judgment about what those patterns mean, and what an organisation is actually ready to do about them, that part hasn't changed. And I don't think it will.
Enterprise UX across logistics, fintech, automotive R&D, healthcare and AI — each project anchored to a business outcome you can verify.
| Client & Industry | Role & Engagement | Discover | Define | Blueprint | Build | Measure | Methods Used | Verified Outcome & Duration |
|---|---|---|---|---|---|---|---|---|
| Syensqo Solvay spin-off · Global Life Sciences | Principal Service Designer Full E2E · sole SD lead · C-suite | Service blueprinting Swimlane mapping Change management Design system | €500K saved 100+ services blueprinted · 24K employees · 8 regions · zero disruption 2023 – Ongoing | |||||
| Newport Tank Containers · Global logistics Logistics B2B | Service Designer Discovery → Blueprint · 6-actor | Journey mapping Multi-actor blueprint Stakeholder interviews Co-design workshops | $2M avoided Demurrage cut · NPS 12→+47 · processing time −40% 9 months · Completed | |||||
| Rappit.io Enterprise SaaS · Low-code Developer Tools | Principal UX Lead Full E2E · team of 4 · C-suite | 25 user interviews WCAG 2.1 AA GenAI integration Lean Six Sigma | 4× velocity $4M revenue · 94% task success · 80K DAU · 50% rework cut 2+ years · Completed | |||||
| Omoda European fashion retail · B.V. Retail & Wholesale | Lead UX Designer Discovery → Measure · ERP replace | AS/400 ERP migration Mobile-first UX Usability testing Change management | +31% productivity Errors 8.3%→2.7% · satisfaction 2.1→4.3/5 · 15-yr legacy replaced 18+ months · Completed | |||||
| eConnect AI invoice automation · NL AI / FinTech | Lead UX Designer AI/ML Discovery → Measure · human-in-loop | Human-in-the-loop UX AI confidence design ML workflow mapping Trust calibration | €1.7M saved AI trust 2.1→8.7/10 · 1.8M+ invoices/yr automated 18 months · Completed | |||||
| Valeo Automotive R&D · SA · Global Automotive R&D | Lead Product Designer Define → Measure · SAP-integrated | SAP integration design Enterprise UX Milestone planning Multi-region rollout | 67% faster setup 22K engineers · 66 R&D centres · 8,000+ concurrent projects 2024 – Ongoing |
Syensqo SA · Principal Service Designer · Ongoing 2023–
A 90-day legal deadline to separate 134 business-critical services from Solvay — across 13,000 employees, 30 countries, ISO and GxP compliance. The insight that drove everything: if new services preserve existing mental models, you don't need a training programme. €500K saved. Zero working days lost.
Newport Tank Containers · Lead Service Designer · 14 months
Depot operators, drivers, customs, agents, brokers, and customers had never shared a service view — each worked in isolation, generating $2M in annual demurrage nobody could attribute. The blueprint made the dependencies visible for the first time. NPS moved from 12 to +47 in 12 months.
Rappit.io · Principal UX Lead · 2012–2025
Developers weren't afraid of AI-generated code. They were afraid of code they couldn't inspect, own, or debug. An always-visible code preview — never hidden, never abstracted — turned sceptics into daily users. 4× dev velocity. 89% 90-day retention vs 34% industry average. 80K DAU.
eConnect · Lead UX Strategist · 12 months
Finance professionals rejected the higher-accuracy model because they couldn't see its reasoning. Confidence heatmaps — showing the AI's certainty field by field — tripled adoption at lower accuracy. Transparency outperformed accuracy. €1.7M annual savings. AI trust score 2.1→8.7/10.
Valeo SA · Lead UX Strategist · 18 months
8 weeks of ethnographic research across 5 countries surfaced the core problem: R&D professionals plan in milestone narrative beats, not dependency chains — but every tool assumed otherwise. Replacing Gantt with a visual milestone timeline got 91% A/B preference. 15,000 engineering hours saved per year.
Autitouch · Lead Service Designer · 8 months
Four actors — child, therapist, parent, diagnostician — each needed a different view of the same moment. The hardest constraint: the child must experience the session as play, never as assessment. Sessions needed for diagnostic signal reduced 3→1. Inter-rater scoring variance eliminated.
Omoda B.V. · Lead UX Designer · 14 months
A previous modernisation was abandoned after 18 months, leaving deep organisational scepticism. 40 hours shadowing warehouse workers across three shifts — not in a conference room — revealed the prior system failed because it was designed for screens, not for people in safety gloves under variable lighting. 67% error reduction. +31% picker productivity.
95%-accurate AI was rejected by finance professionals. Redesigned with confidence heatmaps at 87% accuracy — trust tripled. Transparency outperformed accuracy.
"Tremendous value and time-savings to our thousands of engineers managing more than 8,000 projects globally on a daily basis."
"A core enabler to our ambition to provide a fully automated service. The human-centred approach to AI has been crucial for user adoption."
Verifiable via LinkedIn recommendations · linkedin.com/in/majujmathew
Open to senior design leadership roles, advisory engagements, and strategic consulting. If the role is right, I'll know before the JD finishes loading.
Your details will only be used to respond to your enquiry and will not be shared with third parties.
