Risk Managers & Quants
Daily task
Monitor live market exposure and drill into transaction-level detail across trading desks.
Needs raw density. Full data on demand, zero cognitive overhead.
Case study 02
Decision-Critical by Design
Six years as Founding Lead Product Designer on Compatibl — a financial risk platform serving Central Banks, G15 derivatives dealers, and global asset managers. I led product design across three high-stakes user segments, shipping 30+ enterprise deployments where design defensibility mattered as much as design quality.
01The Context
Six years designing decision-critical tools for people who cannot afford to be wrong.
A white-label financial risk platform serving Central Banks, G15 derivatives dealers, and global asset managers. Every deployment was a fully branded, client-specific product — built on the same underlying system.
Three user types, three workflows, one non-negotiable shared requirement: zero tolerance for error in high-stakes decisions.
Portfolio & Asset Managers
Daily task
Synthesise market positions and execute daily strategy across diverse asset classes.
Needs synthesis. Fast pattern recognition across large data sets.
Compliance Officers
Daily task
Audit transaction trails and enforce regulatory workflows across all deployments.
Needs auditability. Reproducible views, zero ambiguity.
02The Challenge
One platform. Contradictory needs. Regulated environments where design mistakes cost real money.
Three user segments pulling the platform in different directions. One shared, non-negotiable constraint that unites them.
Every design decision on this platform had to serve all three workflows without compromising any of them — and hold up under regulatory audit.
01
Users think differently
A Quant needs raw density, a Portfolio Manager needs synthesis, a Compliance Officer needs auditability. The same screen has to serve all three without compromise.
02
Every client is a bespoke deployment
Unique regulatory context, branding, and workflow requirements. Custom-building for each one wasn't sustainable.
03
Errors are not acceptable
A mislabelled column, an ambiguous state, a delayed refresh can trigger real-world financial consequences. Design has to be defensible under audit.
03What I Did
Built the product design foundation that let a small team ship 30+ enterprise deployments.
A repeatable four-step approach applied across every deployment — from user segment discovery through cross-functional delivery.
The same working framework whether the client was a Central Bank navigating regulatory compliance, or a G15 dealer optimising real-time trading workflows.
01
Discovery
Multi-segment user research
Direct interviews and workflow observation across three user segments. Every design decision traceable to a specific user need.
02
Strategy
Strategic system framing
Every client request had to fit an existing component or become a universal one. Client-specific screens stopped existing.
03
Execution
High-density data workflows
Custom-built analytical tools designed for expert users. Progressive disclosure by default, full density on demand.
04
Delivery
Cross-functional partnership
Direct collaboration with CEO, C-level, and 10+ engineering units. Every architectural decision made with engineering in the room.
04Deep Dive 01 —
The Pivot Table Engine
The Pivot Table Engine
Replacing a rigid developer-built tool with the analytical engine at the centre of the platform.
Analysts were operating in a fragmented loop — export data to Excel, manipulate it manually, rebuild filter configurations from scratch every session. The existing table was built by developers using a third-party framework. It failed under exactly the conditions financial professionals used it: complex filtering, multi-level grouping, real-time updates.
Universal engine, not per-team tables
Tension
Each user segment wanted a table tool tailored to their workflow.
Decision
Build one "Master Engine" that every segment configures independently.
Reasoning
Custom per team meant three codebases to maintain. One configurable engine served Risk Managers drilling into transaction-level exposure, Portfolio Managers synthesising market positions, and Compliance Officers auditing data trails — with zero custom builds.
User workflow: 15 min → 3 min
One analyst quote, matched by consistent feedback across sessions.
Excel dependency eliminated
Saved layouts + drag-to-Excel removed the daily rebuild cycle.
Integrated into every product
Became the analytical engine at the centre of the platform.
1 / 5
![]()
05Deep Dive 02 —
The BI Module
The BI Module
Turning a multi-tool analyst workflow into a single-platform investigation loop.
Analysts producing routine deliverables — including daily risk reports distributed by internal bank mail — operated across three disconnected tools. Data in the platform. Charts in Tableau or Excel. Reports assembled in a third system. By the time the chart was built and the report sent, the underlying data had already updated. Analysts were reporting on information that was stale before it reached the reader.
Chart types organised by analytical intent
Tension
Maximum chart flexibility versus decision paralysis for new users.
Decision
Group chart types by the question the analyst is asking, not by visual format.
Reasoning
Analysts don't think "I want a scatter plot." They think "I want to see how two variables relate." Organising by question reduced configuration confusion without removing any capability.
Multi-tool workflow eliminated
Data, visualisation, and export live in one place on the same real-time data model.
Included in every product built
Adopted as a default component across every deployment, driven by stakeholder demand.
Side-by-side comparison
A direct response to repeated client requests, made native to the platform experience.
1 / 7
![]()
06Key Design Decisions
Three decisions that defined six years of work at platform level.
01
Configuration over customisation
Rejected
Custom per client
→
Chosen
Configure existing components
Tension
Clients wanted bespoke features. Engineering couldn't sustain custom builds.
Decision
Every request had to fit an existing component or become a universal one.
Reasoning
A one-client feature is technical debt disguised as progress. Grow horizontally.
02
Custom UI framework, not third-party
Rejected
Third-party libraries
→
Chosen
Custom-built analytical layer
Tension
Third-party frameworks broke under complex filtering and multi-level grouping.
Decision
Build the analytical UI layer in-house, purpose-designed for financial workflows.
Reasoning
Standard tools fail where financial pros work. Custom was the only stable path.
03
Progressive disclosure by default
Rejected
Simplify for the novice
→
Chosen
Full density on demand
Tension
New users need clarity. Expert users need immediate power.
Decision
Every tool opens in summary state, expands to full detail on demand.
Reasoning
Simplifying for novices removes power experts need. Progressive disclosure serves both.
07Outcomes
Scale that a small team could not have delivered any other way.
10+
Cross-functional partners
Engineering teams partnered with across the ecosystem.
30+
Enterprise deployments
Central Banks, derivatives dealers, and global asset managers.
900+
Screens across the ecosystem
One design language, multiple branded contexts.
15→3
Workflow time reduction
Analyst workflow time on the Pivot Table engine, in minutes.
Same-day Turnaround
Client demos shipped in hours, not sprints
A PM could configure a product-ready demo in hours — assembling existing components without touching design or engineering.
Months → 1 week
Full client rebrand in one week
What previously required months of iterative design was reduced to a single one-week configuration — repeatable across every engagement.
1–3 Designers
Three designers, thirty products, one system
Design capacity scaled through the system architecture — not through hiring. One team, one design language, every product context.
100% Audit pass
Audit-defensible across every Tier-1 client
Design decisions were defensible under review — from Central Banks to derivatives dealers to global asset managers.
"The system didn't just reduce costs. It made commitments possible that would otherwise have been impossible."
08Leadership
The team scaled through the system, not through hiring.
Mentorship
8
Designers mentored into product specialists
From graphic and web backgrounds to mid-level product designers — through structured feedback, portfolio reviews, and career development.
C-level partnership
10+
Engineering teams partnered with directly
Working sessions with CEO and C-level to bridge business direction and engineering capacity — every architectural decision made with engineering in the room.
Ecosystem governance
30+
Enterprise deployments governed as one design ecosystem
Reducing design-to-engineering handoff friction across the full product portfolio — one design language, multiple branded client contexts.
Mentorship trajectory · 8 designers
Starting point
Graphic & Web Designer
Visual craft. Client work. Limited exposure to product systems.
→
Transition
Junior Product Designer
Component thinking. Cross-functional collaboration. Framework fluency.
→
Outcome
Mid-level Product Specialist
System ownership. Stakeholder-facing. Autonomous project delivery.
09What I Learned
01
Expert users don't need simpler — they need respected.
The design brief for a Risk Manager is not "make it easy" — it's "get out of my way." Simplifying for the novice removes power the expert needs. Respecting the expert removes friction the novice never had.
02
Configuration is a design decision, not an engineering one.
The choice to design for reuse instead of custom is architectural — made at the design layer, not inherited from engineering constraints. Every component either scales horizontally or becomes technical debt.
03
Trust is built at handoff, not at pitch.
C-level stakeholders don't judge design by the vision. They judge it by whether it survives contact with engineering, compliance, and clients. Every polished mockup is a hypothesis until it ships.
04
In regulated environments, "good enough" is a legal risk.
Design defensibility matters as much as design quality. A mislabelled column, an ambiguous state, a delayed refresh — each carries real-world consequences beyond user friction.
1 / 7
![]()