KELA Group - Cyber Intelligence

TLDR

My role as the sole product designer, involves implementing diverse research methods, creating & maintaining the design system, and being responsible of end-to-end user experience across the product.

Integral to the product team, I collaborate closely with R&D teams, stakeholders, and various departments, navigating the challenges inherent in our complex system.

Raveed Laeb

VP Product @ KELA

Omer is a true design Swiss Army Knife: he can do marketing, UX, UI, wireframes and copy; he can create detailed, granular, pixel-perfect design systems in the morning and super quick-and-dirty wireframes in the afternoon; and above everything else, he's proactive, pushy (in a good way!), and can make things happen. He was a sole, part-time designer in a messy and agile setting - but knew exactly how to make the most out of it and how to move the needle with great results!

• Product Design
• Research
• Complex System
• Design System
• SaaS
• Marketing
• B2B

Overview - Product designer for the KELA group

In the dynamic landscape of cybersecurity, my role as the sole product designer at KELA was pivotal in reshaping the user experience and overall design strategy across the company's products. Over two years, KELA grew from a single flagship cybercrime intelligence platform into a portfolio - adding UltraRed (external attack surface management) and AiFort (AI red-teaming for LLMs). I designed across all three, owning everything from research and IA to the design system, data viz, and marketing.

What I owned - 6 pillars of impact

1. Complete redesign of the platform

Challenge: The existing products had grown organically and suffered from inconsistent UX, broken navigation, and dated visuals.

Solution & benefits: A comprehensive redesign grounded in user-centric design principles, intuitive navigation, and a cleaner, more modern visual language - driving stronger user engagement and a more positive everyday experience.

2. From three siloed products to one platform

Challenge: KELA shipped multiple products as separate apps - fragmented UX, missing cross-product flows, and a heavy dev burden maintaining each one separately.

Solution & benefits: I unified the products into one cohesive platform - streamlined features, a seamless cross-product user journey, reduced confusion for customers, and dramatically lower maintenance overhead.

3. Design system - build & maintain

Challenge: Design inconsistencies and prolonged dev times were slowing every team down.

Solution & benefits: Built and maintained a comprehensive design system - variants, variables, tokens, and documentation - that powered every KELA product, plus UltraRed and AiFort. It cut development time, accelerated time-to-market for new features, and gave the platform a single confident voice.

4. Cross-functional collaboration

Challenge: Silos between product, engineering, marketing and stakeholders made it hard to ship anything consistent.

Solution & benefits: Daily collaboration with PMs, devs, and stakeholders. I set up collaborative tools, working methodologies, and guidance for the dev team that kept design and engineering tightly aligned through implementation.

5. Introducing research as a practice

Challenge: Before I joined, no structured research existed - every product decision was based on intuition.

Solution & benefits: Introduced and embedded research into the team's workflow - user interviews, usability testing, and surveys - so the team could make decisions grounded in real user insight. I also wrote internal mini-guides (interview script, usability testing) to preserve the practice as the team grew.

6. Brand & marketing design

As the sole designer, I also owned all marketing-side work: the company website, datasheets, landing pages, webinar artwork, social graphics, and print collateral.

Case study - Monitor Alerts: rebuilding a core product

The Monitor Alerts product is one of KELA's core modules - a system for managing the threat alerts the platform surfaces. Each alert can carry wildly different content, there are dozens of alert types, and the legacy implementation had paid for that flexibility in UX debt: powerful but cluttered, with known UX pain (per customer feedback) and a maintenance load that bred bugs. I led the full rebuild.

Stage 01 - Requirements

Rebuild a core (legacy) product from the ground up. Simplify UX. Rethink user flows and the way Monitor connects to the rest of the platform. Move to the new design system. Because Monitor sits at the center of what customers actually buy, the cost of a wrong call was high - so I committed to grounding every decision in user research wherever I could fit it in.

Stage 02 - User Research

Wrote an interview script grounded in NNG best practices (no leading questions, no past-tense or predictive questions), interviewed internal analyst users, and synthesized common pain points and patterns.

Bonus: I turned what I learned into a mini-guide so the small product team could keep running the same kind of research after I'd handed it off.

Stage 03 - UX, product & visual inspiration

With research insights in hand, I gathered references on two axes: how competitors and adjacent systems structured this kind of data (hierarchies, actions, surfacing critical info), and the visual side (page layouts, key components, forward-looking thinking on data viz for downstream modules). I looked for recurring patterns to establish a baseline of convention, then used them to defend the choices I proposed to the product team.

Stage 04 - Mockups & Prototypes

Combining research insights, the product team's accumulated knowledge, and the competitor research, I built two distinct layout concepts - each solving the problem in a different way - then ran the product team through pros and cons before picking one.

Table layout: denser, more rows per screen, conventional. But Monitor's records can themselves contain tables - a table inside a table felt heavy, and 'simpler & more intuitive' was the brief.

Split-pane (selected): familiar pattern (Gmail / Slack / WhatsApp), an alert is always selected so its details are open in the same view. Less context-switching, easier scan. Trade-off: bulk actions are a little less efficient. Worth it.

Stage 05 - Usability Testing

Wrote the test script (again grounded in NNG best practices), ran it with the same internal users I'd interviewed, compared results, found recurring failure patterns, and iterated the mockups until they cleared the bar.

Bonus: a second mini-guide - this one for running usability tests - went into the team's playbook.

Works - Data visualization across the platform

Beyond Monitor, I designed the data viz language used across KELA's modules - favoring shapes that read fast at a glance and degrade gracefully when the underlying data is sparse, noisy, or critical.

APTs Dashboard

An interactive map of advanced persistent threat groups, linking attacker origins to targeted geographies - so analysts can see, at a glance, who's hitting whom.

Actors Hub

A mini-dashboard summarizing the top valuable stats on threat actors - targeted sectors, geographies, motivations - while keeping a fast filter on the feed below.

Identity Guard

Same compact-dashboard pattern, applied to compromised-identity intel: incidents created today, severity split, service-type distribution, threats by incident, and geography.

UltraRed - external attack surface management

Vectors

The Vectors module surfaces individual attack vectors and their priority. I designed a custom gauge graph to make the priority score immediately readable inside dense list views and detail pages.

Platform Dashboard

The UltraRed platform dashboard rolls up the main stats from every critical module (MTTR, vector score, asset scores) so customers can size up their attack surface in a single glance - and dive into the right module when something needs attention.

AiFort - AI red-teaming for LLMs

Models Overview

The main dashboard for an AI model under test - surfacing the most relevant data, score trends, and per-category insights, plus the list of test plans currently running against it.

Model - Test Plan

Drill-down view of a single test plan, highlighting the relationship between questions and prompts via a category-scored heatmap.

Model tests

A high-level benchmark view across AI models - side-by-side scores and trend lines over time, so customers can compare ChatGPT, PaLM, Gemini and others on the same canvas.

Outcome - From single product to platform group

Over two years, KELA's product design function went from 'nice-to-have' to a multiplier - driving a redesigned, consolidated platform, an efficient design system, embedded research, and the data-viz language used across every module of KELA, UltraRed, and AiFort. The work helped position KELA as a leader in cybercrime threat intelligence, with a user-centric approach that continues to compound across the group's products.