Team | Solenya

Our Team

A blend of the finest ingredients to always hit the spot.

The Solenya team

From machine learning researchers to engineers, we're all fermenting ideas into reality—one dill-icious innovation at a time.

Karl Hammerschmidt

Marcus Gawronsky

Cumin Taste Optimizer

github

Nicholas Wilkinson

Mustard Level Engineer

scholar

Jean Durand

Preston Govender

Mystic Laser Engineer

github

Lawrence Godfrey

github

Amandla Mabona

Matzo Larder Researcher

scholar

Simon Innes

Dolmas Engineer

Miguel Da Costa

Company Onion Organiser

Our Offices

Two timezones, one kitchen. London ships product, Cape Town ships papers.

Blogs

Open-notebook research. Methods, confidence intervals, and the experiments that didn't work.

Beyond Average Revenue: A Four-Part Decomposition for Measuring Discovery Experiments

Beyond Average Revenue: A Four-Part Decomposition for Measuring Discovery Experiments

Mean revenue is a noisy metric for A/B tests in e-commerce - it conflates purchase probability, visit frequency, and basket size into a single estimand. We decompose them, fit each with an appropriate GLM, and embed the result inside a 2² factorial design to measure two platform features simultaneously. The approach recovers clean per-feature estimates even under a confounded allocation mechanism.

SOL-BRN-024In Stock · 12 min read

Architecting Solenya API V1 for the Agentic Era

Architecting Solenya API V1 for the Agentic Era

Our new API documentation is live. This is the lab notebook on the philosophy behind every design decision-from progressive disclosure and GraphQL introspection to sendBeacon analytics and OIDC delegation-and the single architectural thread that connects them all.

Organic AICrunchy

SOL-BRN-023In Stock · 14 min read

The Arrival of the Cognitive Core: Why Tools Are Now Everything

The Arrival of the Cognitive Core: Why Tools Are Now Everything

Andrej Karpathy predicted the optimal reasoning model would be lean, stripped of rote memory, and reliant on tools. With Qwen 3.5 4B, that prophecy is fulfilled-and it means the entire competitive landscape pivots to tool infrastructure.

Extra DillFermented

SOL-BRN-022In Stock · 12 min read

The Three Waves of Agentic Tools: From API Wrappers to Self-Optimizing Sub-Agents

The Three Waves of Agentic Tools: From API Wrappers to Self-Optimizing Sub-Agents

Tracing the evolution of tool use in AI: from legacy API wrappers to specialized Small Language Models, and the explosive rise of self-optimizing sub-agents driven by harsh hardware realities.

SOL-BRN-021In Stock · 10 min read

The Agentic Protocol Ecosystem: How MCP, A2A, and Commerce Protocols Are Reshaping the Web

The Agentic Protocol Ecosystem: How MCP, A2A, and Commerce Protocols Are Reshaping the Web

A deep dive into the emerging protocol stack-from MCP's founding to the ACP vs UCP aggregator wars-and how these layers are turning static storefronts into negotiating sub-agents.

SOL-BRN-020In Stock · 10 min read

The State of Embedding-Based Retrieval in E-Commerce

The State of Embedding-Based Retrieval in E-Commerce

We explore the current state of embedding models for search, why they fall short in e-commerce applications, and the industry's duct-tape solutions.

Organic AICrunchy

SOL-BRN-019In Stock · 10 min read

The One-Model Approach: A Paradigm Shift in Product Discovery

The One-Model Approach: A Paradigm Shift in Product Discovery

When your architecture is fragmented, you're not just adding complexity, you're encoding fundamental limitations into your entire stack. Here's why a unified model approach creates cascading benefits far beyond the obvious.

SOL-BRN-018In Stock · 8 min read

Users and Items: Rethinking E-Commerce User Modeling

Users and Items: Rethinking E-Commerce User Modeling

While CLIP revolutionized product representation, user modeling has been an afterthought. We explore why treating users and items symmetrically fails to capture the temporal, uncertain, and compositional nature of shopping intent-and what's needed to bridge this gap.

Zero-shotExtra Dill

SOL-BRN-017In Stock · 6 min read

Composed Item Retrieval: A Whole New World

Composed Item Retrieval: A Whole New World

Explore how Composed Item Retrieval is going to change e-commerce by allowing shoppers to modify product searches with natural language, creating a more intuitive and powerful discovery experience.

Zero-shotExtra Dill

SOL-BRN-016In Stock · 5 min read

It's All Multimodality

It's All Multimodality

Time to let the flavours blend together... For too long, e-commerce discovery has been a messy kitchen of separate tools for text, images, and recommendations. We trace the recipe from plain-toast keywords to today's complex dishes, and show why the future is a unified kitchen that can serve up truly novel shopping experiences.

SOL-BRN-015In Stock · 6 min read

E-Commerce is Inherently Multimodal

E-Commerce is Inherently Multimodal

Don't trust the headlines, CLIP is not multimodal! In this blog post, we explore the multimodal nature of e-commerce and how it can be leveraged for better user experiences.

SOL-BRN-014In Stock · 4 min read

Sizing Zero-shot Recommendations

Sizing Zero-shot Recommendations

In this blog post, we explore the cold start problem in e-commerce recommendations, quantifying missed opportunities with traditional systems and showcasing the potential of zero-shot recommendation approaches.

Organic AICrunchy

SOL-BRN-013In Stock · 6 min read

Smarter Sorting, More Sales: Evaluating the Impact of Fixed Sorts

Smarter Sorting, More Sales: Evaluating the Impact of Fixed Sorts

The way you sort products has a direct impact on your bottom line. This blog dives into a simulation that shows how a more dynamic approach can prevent premature stock-outs and ultimately drive higher cumulative sales.

SOL-BRN-012In Stock · 6 min read

Models are Markup, Tokens are Features...

Models are Markup, Tokens are Features...

We explore the interplay of token architecture on product design, and model development. We discuss how tokens can be used to create a more capable, flexible and adaptable AI systems.

ArtisanalMultimodal

SOL-BRN-011In Stock · 6 min read

The Latest Squeeze on Training Bottlenecks...

The Latest Squeeze on Training Bottlenecks...

We unravel the latest squeeze on training bottlenecks, exploring the impact of data quality and pipeline monitoring on training performance.

Organic AICrunchy

SOL-BRN-010In Stock · 14 min read

GPU Poor, GPU Rich: Training at Scale

GPU Poor, GPU Rich: Training at Scale

We go down the rabbit-hole, exploring the recipes that train transformers at scale. These provide insight into network communication, GPU utilization, and the challenges of common transformer ops.

ChunkyOrganic AI

SOL-BRN-009In Stock · 14 min read

Architecture & Optimising: Determining Training Memory

Architecture & Optimising: Determining Training Memory

Understand the recipe and you'll know what comes next! We break down the memory requirements of transformer training and weigh in on some innovations born of these constraints.

Zero-shotExtra Dill

SOL-BRN-008In Stock · 14 min read

Low-Level Performance Optimization in PyTorch on NVIDIA GPUs

Low-Level Performance Optimization in PyTorch on NVIDIA GPUs

Delve into the low-level details of Nvidia GPUs and their interaction with PyTorch execution.

SOL-BRN-007In Stock · 17 min read

Tracking Model Experiments

Tracking Model Experiments

Let's jump into the Experiment Tracking - a guiding light in large-scale model training.

SOL-BRN-006In Stock · 10 min read

Vector Databases: A New Dimension in Data Storage

Vector Databases: A New Dimension in Data Storage

Dive into the technology behind vector databases, one of the key innovations helping drive modern AI.

Extra DillFermented

SOL-BRN-005In Stock · 8 min read

The Impact of Zero-shot Recommendations

The Impact of Zero-shot Recommendations

Find out how Zero-shot Discovery unifies the search and recommendations experience in e-Commerce, and the impact legacy algorithms have on user growth.

Organic AICrunchy

SOL-BRN-004In Stock · 5 min read

What is Zero-shot Discovery? Let's dive in.

What is Zero-shot Discovery? Let's dive in.

What makes this wave different? Here are how economics and hardware are building innovation in AI and Machine Learning that will last.

MultimodalSpicy

SOL-BRN-003In Stock · 3 min read

The Five Technologies driving AI 3.0

The Five Technologies driving AI 3.0

What makes this wave different? Here are how economics and hardware are building innovation in AI and Machine Learning that will last.

Organic AICrunchy

SOL-BRN-002In Stock · 6 min read

In a Pickle About AI? Let's Relish the Latest Developments!

In a Pickle About AI? Let's Relish the Latest Developments!

Crunch time in AI research! Our tangy tour through the latest breakthroughs in PeFT, Upcycling and Zero-shot Learning will leave you feeling anything but sour.

Zero-shotExtra Dill

SOL-BRN-001In Stock · 2 min read

Verified by