Product designer with a background in architecture. Three years in climate tech, designing for complex domains where making information feel manageable is the hard part.
I use AI tools daily to move faster: quicker prototyping, tighter HTML/CSS handoff with engineering, more time on research and design decisions.
CV
Six projects. Three years. One platform that finally scaled.
The Field Management Page and Cover Crop Mixer didn't come from a brief. They came from three years of shipping things that didn't quite work, learning why, and trying again. Six projects. Each one exposed something the last one got wrong.
[Focus]Platform architecture, design system, iterative strategy, cross-functional collaboration, design leadership
[Timeline]3 years, 2022–2025
[Platform]Web & Mobile
2022
PROJECT 00
Original Mobile Platform Data entryInheritedDeprecated
Pre-2022
WHAT IT WAS
A mobile-first platform designed by Agnieszka Sowinska before I joined. Fields were cards, with regenerative practice modules (tillage, fertiliser, crop rotation, plant protection) assigned to each one. Mobile app first, web second.
ITS LIMITS
Card-based fields didn't scale well. The mobile-first assumption also turned out to be wrong: farmers used this at a desk, managing many fields at once, not on a phone in a field. And there was no way to simulate payouts yet.
DESIGN
Agnieszka Sowinska (Lead)
PRODUCT
1 PM
ENGINEERING
2 engineers
I join → first project: Benefit Calculator
PROJECT 01
Benefit Calculator SimulationDeprecated
Q3 2022
WHAT IT WAS
My first project at Klim. A simple tool for farmers to estimate potential carbon payouts before committing to documentation, built on top of the existing mobile platform. The idea was to make the value of regenerative practice tangible early.
WHAT IT REVEALED
The crop comparisons were misleading. The section on results was not highlighted enough. The user journey didn't hold together. It exposed how much the underlying simulation model needed work, and got me thinking about timeline-based planning.
DESIGN
Antoine Richard (Solo designer)
PRODUCT
1 PM
ENGINEERING
2 engineers
Insight → simulation needs to be more rigorous
2023
PROJECT 02
Platform Rebranding + Farm Page Redesign Data entryDeprecated
Q1–Q2 2023
WHAT IT WAS
Fields moved from mobile cards to desktop tables, and the visual identity got a full redesign to convey a more professional interface. This was when the team committed to desktop-first, because that's actually how farmers planned their work.
WHAT IT REVEALED
The mobile-first assumption was just wrong. Farmers sat at desks. They managed many fields. They planned ahead. Desktop wasn't a nice-to-have; it was the real surface. But the platform was still a loose collection of tools that didn't talk to each other.
DESIGN
Agnieszka Sowinska (Lead), Antoine Richard
PRODUCT
2 PMs
ENGINEERING
8 engineers
Insight → desktop first, simulation needs a timeline
PROJECT 03
Crop Explorer SimulationRecommendationDeprecated
Q4 2023
WHAT IT WAS
I led the design of the Crop Explorer, a feature for planning regenerative practices across multiple years, tied to carbon credit outcomes. The first time we introduced a real timeline into the product, and the first attempt at multi-year rotation-based simulation.
WHAT IT REVEALED
Farmers responded well to timeline-based planning. Seeing projected payouts over multiple years was engaging in a way earlier tools hadn't been. But the Crop Explorer was still its own island, another standalone tool adding to the platform's fragmentation problem.
DESIGN
Antoine Richard (Solo Designer)
PRODUCT
1 PM
ENGINEERING
4 engineers
Q&A
1 QA
Insight → timeline works, but standalone tools don't
2024
PROJECT 04
Digital Agronomist Tool Concept prototypeSimulationRecommendation
Q3 2024
WHAT IT WAS
A concept prototype I led, focused on long-term crop rotation and agronomic recommendations. It also set up the timeline-based planning logic that later became central to the Field Management Page.
THE BREAKING POINT
Cover crops shouldn’t be treated as single species, but rather mixes. Every standalone tool added implementation cost and inconsistency. Farmers were navigating four different interfaces with four different logics. The cognitive overhead was real. At some point it became obvious: we couldn't keep adding tools. The whole architecture needed rethinking.
DESIGN
Antoine Richard (Solo designer)
PRODUCT
1 PM
AGRONOMY
2 agronomists
DATA
1 data scientist
Breaking point → four siloed tools → platform redesign
PROJECT 05
Field Management Page FoundationCurrentData entryData planningRecommendationSimulation
Q4 2024
WHAT IT WAS
A turning point for the platform. New carbon compliance requirements, updated calculation logic and program rules (from inventory and intervention to footprinting), and the need to integrate recommendations directly into the data entry flow made a full redesign necessary. The timeline logic, the recommendation model, the field-first structure: ideas I had explored in the Crop Explorer and Digital Agronomist finally came together in a single unified experience.
WHAT IT ENABLED
An architecture that could actually grow. One interface for documenting various practices and carbon programs, running simulations, and getting recommendations. And a stable enough foundation that new tools like the Cover Crop Mixer could be built on top of it without breaking everything else.
DESIGN
Tim Schober (Lead), Antoine Richard
PRODUCT
up to 5 PMs
ENGINEERING
12 engineers
DATA
3 data scientists
Q&A
4 QA
Foundation → integrated tools now possible
2025
PROJECT 06
Cover Crop Mixer CurrentData entryData planningRecommendationPersonalisationSimulation
Q1 2025
WHAT IT WAS
I led the design of the Cover Crop Mixer, a goal-driven tool for building personalised cover crop mixes. It lives inside the Field Management Page and translates agronomic species rules into a guided four-step flow. The first thing we built that fully used the new platform architecture.
DESIGN
Antoine Richard (Solo designer)
PRODUCT
1 PM
ENGINEERING
3 engineers
AGRONOMY
2 agronomists
DATA
1 data scientist
Q&A
1 QA
Learnings
Every project taught us something the next one needed. Not in a tidy way. Each one hit a wall, exposed something wrong, and forced a rethink. The timeline came from the Crop Explorer running out of road. The recommendation model came from the Digital Agronomist showing us what standalone tools couldn't do. The desktop-first shift came from realising the mobile assumption was just wrong. None of it came from a brief. It came from getting things wrong first. That's what made the Field Management Page possible.
More work