From Team Augmentation to Product & AI Partnership for a Traveltech Leader
Ongoing partnership delivers software modernisation, product innovation, and AI activation for the CarTrawler engineering team.
Executive Summary
CarTrawler is a global B2B travel technology provider powering car rental and mobility experiences for hundreds of airlines and travel brands worldwide. Operating at massive scale, the company processes over 20 million daily searches across a network of more than 1 billion vehicles.
The Problem
In 2021, CarTrawler partnered with Zartis to help solve an immediate challenge: scaling high-quality engineering capacity in a competitive hiring market while modernising a complex, business-critical frontend platform.
The Solution
What started as a single Zartis engineer embedded in a CarTrawler squad has evolved into a long-term, multi-discipline partnership. As of October 2025, the team has grown to 15 specialists across engineering, product, QA, and Android, and now also includes strategic product and AI support from Zartis leadership.
The Outcomes
- Re-architected a core booking frontend on modern Angular, with modular architecture, improved performance, and dramatically faster delivery.
- Reduced technical debt by rebuilding key frontend surfaces from scratch, keeping pace with major Angular releases up to Angular 20 (with Angular 21 planned).
- Enabled high-profile partner launches such as Ryanair and Southwest in InPath — that would not have been feasible on the previous stack.
- Introduced eventing and telemetry foundations, improved bundle and asset outputs, and performance-critical features like Prewarm for significantly better load times.
- Supported CarTrawler’s emerging AI strategy via a strategic workshop and ongoing lightweight AI experiments focused on real, day-to-day impact.
- Strengthened product organisation effectiveness by embedding Zartis Head of Product, Luis Torregrosa, into CarTrawler’s product team.
CarTrawler’s end-to-end technology platform expands our airline and travel partners’ offering to their customers, creating substantial ancillary revenue opportunities. They provide unrivalled breadth and depth of content worldwide, including car rental, private airport transfer and ride-hailing services.
CarTrawler creates innovative, data-led solutions for some of the largest travel brands in the world, including United Airlines, American Express, easyJet, Alaska Airlines, SWISS, Hotels.com and Emirates. As a B2B company, they focus on helping our airline and travel partners build their brands.
The Business & Technical Challenge
CarTrawler operates in a space where small improvements in conversion rate translate directly into meaningful revenue for both CarTrawler and its partners. Therefore, latency, reliability, and flexibility are critical for partner experience and end-users across the globe.
As the platform evolved and partnership demands increased, CarTrawler faced several pressures:
Growing product complexity
Increasing need to manage complexity across multiple SmartBlock versions and customer experiences
Legacy frontend implementations
E.g. sbip V3 on AngularJS – that slowed development and were becoming harder to maintain.
Competitive hiring market in Dublin
Difficulty in bringing in enough senior engineers with the right experience.
What CarTrawler Needed
On the technology side, CarTrawler needed to:
- Move away from AngularJS and legacy implementations to a modern, scalable Angular architecture.
- Improve performance and load times, especially under heavy load and in demanding booking flows.
- Reduce firefighting and tech debt, freeing teams to focus on new features rather than constantly patching legacy code.
- Align architecture across SmartBlock 4 (SB4) and sbip V3 while preparing for future growth.
At the same time, CarTrawler wanted to understand where AI could realistically help – not as a buzzword, but as a set of practical tools to improve workflows, search and booking quality, and internal decision-making.
Why Traditional Approaches Weren't Working
Slow hiring cycles
Senior engineers with strong frontend and product-scale experience are scarce and competitive.
Limited flexibility
Internal hiring doesn’t easily allow ramping capacity up or down as product priorities shift.
Strategic gap
Staff augmentation alone rarely brings the architectural, product, and AI perspective needed to re-shape a core platform.
Tech debt drag
Without dedicated focus, teams risk spending most of their time keeping legacy systems alive instead of building the next generation.
Why They Chose Zartis
CarTrawler needed a partner who could combine senior engineering execution with architectural leadership and product thinking, embedded within their teams:
- Part of CarTrawler squads, attending standups, sprint ceremonies, and planning sessions.
- Contributing to architecture decisions and technical strategy, especially around SmartBlock 4 and sbip V3.
- Pairing with CarTrawler engineers to share knowledge and best practices.
- Collaborating with product, design, and other stakeholders to align technical work with business goals.
how we worked together
The Zartis Approach
Engineering, AI, and Product Experts Working across Multiple Products
Partnership Growth & Model
- 2021: Engagement begins with 1 engineer, embedded in CarTrawler’s team.
- 2021–2024: Gradual expansion as trust and impact grow; Zartis becomes a key contributor to core frontend work and system evolution.
- 2025: Major expansion with 12 additional engineers, bringing the Zartis team to 15 members as of October 2025.
- Zartis Head of Product, Luis Torregrosa, joins CarTrawler’s product organisation as a strategic partner across key initiatives and AI.
Team Composition
As of October 2025, the Zartis team working with CarTrawler includes:
- 1 × Senior Principal Engineer
- 1 × Head of Product
- 1 × Senior Product Owner
- 1 × Senior QA Engineer
- 1 × Senior Full-Stack Engineer
- 3 × Senior Java Engineers
- 7 × Frontend Engineers
- 1 × Senior Android Engineer
This mix allows Zartis not only to execute, but to shape architecture, product decisions, and delivery practices end-to-end.
Our approach
What We’ve Built & Influenced
Core Frontend Platform: SmartBlock 4 & sbip V3
Zartis has been heavily involved in the SmartBlock 4 (SB4) and sbip V3 frontend platforms – the engines behind critical partner booking experiences. Key architectural decisions and implementations led by the Zartis team include:
Full Modular Architecture Rework
- Re-designed SB4 into a modular structure that improves maintainability, reusability, and testability.
- Enables teams to introduce new features or variants without putting the entire platform at risk.
AngularJS → Angular Migration & Modernisation
- For sbip V3, the team opted for a clean rebuild on Angular, rather than incrementally patching legacy AngularJS.
- SB4 has been kept fully up to date, progressing from Angular 16 to 17, 18, 19, 20, and with an upgrade to Angular 21 planned within weeks.
- Result: A modern frontend that is significantly easier to extend, test, and maintain, with virtually no inherited tech debt.
Performance Features
- Designed and implemented Prewarm and other optimisations with massive load-time impact, improving user experience and partner satisfaction at scale.
- Introduced bundle and asset optimisation strategies (config inheritance, language pipelines, etc.) that reduce weight and improve perceived performance.
Event System & Telemetry Foundations
- Implemented an event system and telemetry layer to give CarTrawler better insight into user behaviour, performance, and platform health.
- This foundation is key for ongoing optimisation, experimentation, and future AI-driven improvements.
SDLC & Delivery: From Weeks to Hours
Improvements made to the SDLC had great impact on development velocity:
Example: A feature that previously took two weeks to implement (on sbip V3) and was prone to issues. The same feature was later implemented on SmartBlock 4 in roughly four hours, with a more stable outcome.
What SB4 enabled:
- Drastically faster iterations
- Safer deployments
- Easier maintainability
- Less firefighting, more product work
More importantly, SB4’s robustness and flexibility made it possible to deliver major partner rollouts:
- CarTrawler team has stated that Ryanair delivery would not have been possible on the legacy v3 approach.
- Ryanair and Southwest in InPath were both made possible by SB4’s new architecture and delivery model.
Work on the React / Monorepo Ecosystem
While the main focus has been on Angular-based platforms (SB4, sbip V3), Zartis engineers have also contributed to the React / monorepo environment:
• Participation in shared component development and integration work.
• Support on customer portal and upsell email pages, where React plays a role.
• Ensuring architectural decisions in Angular-based SB4 remain compatible with the broader multi-framework ecosystem.
Operations & Delivery Changes
Previously, CarTrawler ran a more dynamic deployment model with several releases per week. With the maturity and reliability of SB4 and supporting processes:
• The organisation has been able to move to a target of one deployment per week, at CarTrawler’s request.
• This provides a more controlled, predictable release cadence, without sacrificing the ability to deliver.
Tooling & Infrastructure Touchpoints
While Zartis has not led core cloud migration or infrastructure architecture, the team has:
• Adapted Jenkins pipelines and jobs to align with CarTrawler’s migration to AWS.
• Ensured frontend delivery and build processes are compatible with evolving infrastructure.
The focus remains on the application layer and delivery flow, aligned with CarTrawler’s broader platform strategy.
Our approach
AI Strategy & Product-Level Support
Zartis partnered with CarTrawler to run a strategic AI workshop, focused on:
Beyond engineering, Zartis has strengthened CarTrawler’s product development through the involvement of Luis Torregrosa (Head of Product at Zartis), who is directly embedded in CarTrawler’s product initiatives.
Key areas of support include:
- Running small, low-risk experiments such as the pt-queries GPT proof-of-concept that summarises or classifies partner queries.
- Focusing on quick wins that improve day-to-day workflows without needing large, multi-year programmes.
- Helping CarTrawler teams understand what AI can realistically help with, before committing to bigger initiatives.
- Building a structured, shared view of product capabilities across engines, SmartBlock versions, SDKs, microsites, loyalty, and partner models.
- Organising the framework and gathering inputs from Product & Design, so teams can clearly see what exists today and how features differ across contexts.
- Helping consolidate assumptions, technical findings, business rules, and open questions from multiple teams.
- Ensuring everyone shares the same understanding of scope and uncertainties before estimates or decisions are made.
the results
Outcomes & Impact
Development Velocity & Time-to-Market
From weeks to hours:
Features that previously required weeks now take hours or days on SB4, thanks to modular architecture and improved tooling.
- Faster, more reliable delivery enabled critical partner launches (e.g. Ryanair, Southwest in InPath) that would have been extremely difficult, or impossible, on the previous stack.
- The platform now supports a deliberate weekly deployment cadence while maintaining high velocity internally.
Technical Debt & Platform Stability
Minimised technical debt:
SB4 and the rebuilt Angular frontend represent a full frontend reboot with minimal inherited tech debt.
- Teams moved from a world of constant firefighting on v3 to a more stable SB4 platform where they can focus on new features and partner value.
- Continuous Angular upgrades (16 → 20, with 21 imminent) ensure the platform stays current and supportable.
Product & AI Readiness
AI in Production:
CarTrawler now has a clearer AI strategy, grounded in actual product and engineering realities.
- Lightweight experiments (like pt-queries GPT) give teams concrete examples of where AI can meaningfully improve workflows.
- The Feature Menu and structured discovery work (e.g. AXA → Cover Genius) improve cross-team alignment and decision-making.
Performance & Observability
Faster load times:
- Architecture changes and features like Prewarm deliver significant improvements in load times and perceived performance.ght experiments (like pt-queries GPT) give teams concrete examples of where AI can meaningfully improve workflows.
- The event & telemetry system gives CarTrawler better visibility into user behaviour and system health, enabling data-driven optimisation.
What this enabled next
A Growing Partnership Built for the Future
What began as a single Zartis engineer in 2021 has grown into a 15-person, multi-discipline partnership spanning senior engineering and architecture, product ownership and Head of Product support, QA, Android, and full-stack development, plus strategic AI and product discovery capabilities.
Together with CarTrawler, Zartis has helped modernise a core frontend platform, increase delivery velocity whilst improving stability, and create the foundations for ongoing performance, observability, and AI-driven innovation.
This partnership continues to evolve, with both teams working together to unlock new opportunities in travel technology and deliver enhanced experiences for CarTrawler’s customers and partners worldwide.
Scale Your Engineering & Product Capabilities Without Losing Quality
If you need to modernise legacy systems without pausing feature delivery, nicrease engineering throughput with senior, embedded talent, or bring product and AI thinking into the core of your technical decisions, we can help.