Join us at Experiment 2025

The bottleneck isn't building.
It's knowing what works.


The experiment orchestration layer for AI teams. Remyx learns your codebase, recommends what to try next, and turns every result into context for the next decision.

Start Now
Watch a demo
Featured by the AI community
Remyx — Lifecycle Section Preview
The Lifecycle

AI development is an empirical discipline

Remyx operationalizes the scientific method for AI engineering, managing the full cycle so every stage is instrumented and every decision is preserved.

Observation 01 Hypothesis 02 Experiment 03 Analysis 04 Conclusion 05 Communication 06
01
Observation
Analyze the codebase and survey prior experiments to surface what's worth investigating.
02
Hypothesis
Recommend approaches grounded in your system's history and recent advances.
03
Experiment
Coding agents and your dev stack ship the change. Remyx keeps the provenance.
04
Analysis
Metrics flow in from MLflow, W&B, and your A/B tools, automatically tied to the hypothesis.
05
Conclusion
Record a structured decision (ship, iterate, or reject) with rationale attached.
06
Communication
Roll out, fork a new hypothesis, or prune the search.
How it works

Idea to Production, Systematically

Experiment with confidence.
Integrate discovery, building, and validation.

Email Classification - AI X+ Webflow Template

Hypothesize


Get recommendations grounded in your codebase, experiment history, and recent advances.

Integrated with More Apps - AI X+ Webflow Template

Implement & validate


Coding agents and your tools do the work. Remyx ties every metric, commit, and ticket back to the hypothesis it tested.

Automated Reports - AI X+ Webflow Template

Decide & iterate


Capture the decision and why you made it. Each outcome, positive or negative, narrows the search.

Remyx — Integrations Section Preview
Integrations

Works with your stack

Remyx connects the tools your team already uses into a single experiment record. More integrations ship every month.

Plan & ship
Where changes get planned, shipped, and reviewed
and more
Measure & learn
Where offline and online results get measured
and more
Build & run
Where experiments get implemented and executed
and more
How Remyx Helps

Learn what works.

Remyx helps engineers test more ideas and helps leads know which ones drive improvements.

For AI Engineers

Remyx understands your codebase from day one. Spend your time running experiments, not piecing together what the team already knows.

Email Classification - AI X+ Webflow Template
Searching for what to try next →
Hypotheses recommended from evidence
Career Growth - AI X+ Webflow Template
Context switching across tools →
One integrated flow
Reset Password - AI X+ Webflow Template
Rebuilding context every cycle →
Prior experiment context surfaces automatically

For Team Leads

A portfolio view across every active experiment. Every decision, trajectory, and pending call is visible without interrupting your team's flow.

Get Results - AI X+ Webflow Template
Status spread across Slack and 1:1s →
A single portfolio view
Career Growth - AI X+ Webflow Template
Decisions you can't reconstruct later →
Structured disposition on every call
Reset Password - AI X+ Webflow Template
Tribal knowledge scattered across tools →
A single source of truth
Know What's Next

Each experiment lends more evidence to refine your next hypothesis

The space of possible improvements is too large to try everything. Remyx helps your team start from relevant prior work and build on what you learn.

EXP 1 First hypothesis EXP 3 Narrowing the search space EXP 10 Shared decision record
First hypothesis
You find a technique relevant to your problem. Remyx sets up a pre-built environment so you can test it against your application directly, instead of rebuilding it from scratch.
Narrowing the search space
Your past results inform what to try next. Your next experiment follows from evidence, from what your team has already learned, so you're building on real results, not reacting to hype or following hunches.
Shared decision record
Every technique tested, every result measured, every dead end recorded. New team members onboard from real experiment history. Leadership reviews the reasoning behind decisions alongside the outcomes.
EXP 1
First hypothesis
You find a technique relevant to your problem. Remyx sets up a pre-built environment so you can test it against your application directly, instead of rebuilding it from scratch.
EXP 3
Narrowing the search space
Your past results inform what to try next. Your next experiment follows from evidence, from what your team has already learned, so you're building on real results, not reacting to hype or following hunches.
EXP 10
Shared decision record
Every technique tested, every result measured, every dead end recorded. New team members onboard from real experiment history. Leadership reviews the reasoning behind decisions alongside the outcomes.

Every experiment is a decision: a hypothesis, a result, a rationale.
Experiment orchestration keeps that record compounding instead of evaporating, so your codebase evolves on evidence.

Learn how ExperimentOps works
From the Founders

Built by practitioners, for teams building at the frontier

A team of mathematicians and award-winning ML innovators with a decade of experience applying AI in robotics, healthcare, content recommendation, and enterprise data/ml infrastructure.

Salma Mayorquin
Salma Mayorquin
CEO & Co-Founder

Applied Mathematics, UC Berkeley. Former Solutions Architect at Databricks advising MLOps strategy from startups to Fortune 500. Award-winning ML innovator recognized by NVIDIA's developer community.

Enterprise MLData InfrastructureAI Strategy
LinkedIn →
Terry Rodriguez
Terry Rodriguez
CTO & Co-Founder

UC Berkeley. 10+ years applying ML in healthcare, robotics, and content recommendation at Riot Games, Tubi, Robust.AI. Open-source tools cited by Google DeepMind and used in peer-reviewed research.

RoboticsHealthcare AIContent Rec
LinkedIn →

Talks, Pods & Writing

Conference talks, podcast conversations, and field notes on how AI teams go from experiment to production.

🎙Podcast · Heavybit
ExperimentOps: Why MLOps Alone Isn't Enough
Salma unpacks the shift from MLOps to ExperimentOps with stories from Netflix, Yelp, and Stripe on bringing foundation models into production.
Salma MayorquinListen →
🎤Talk · MLOps Community
Beyond Benchmarks: Measuring Success for AI Initiatives
Why offline metrics don't predict online success and how to design evaluation systems that align model development with business outcomes.
Salma MayorquinWatch →
⚡Talk · AG2 Community
GitRank: Automating Paper-to-PR with Agentic Code Execution
Technical deep-dive on how GitRank uses AG2 agents to discover, build, and test research implementations from arXiv to live experiments.
Terry RodriguezWatch →
📰Feature · Cerebral Valley
Remyx: Your AI Production Assistant
In-depth interview on the mission behind Remyx, how the founders met at Berkeley, and their vision for the next generation of AI engineering tools.
Salma MayorquinRead →
✍️Blog · Substack
Vibes Don't Scale
Why AI-assisted coding falls apart past the demo stage and what disciplined experimentation looks like when complex code changes are under-determined by simple prompts.
Terry RodriguezRead →
🎙Podcast · Heavybit
ExperimentOps: Why MLOps Alone Isn't Enough
Salma unpacks the shift from MLOps to ExperimentOps with stories from Netflix, Yelp, and Stripe on bringing foundation models into production.
Salma MayorquinListen →
🎤Talk · MLOps Community
Beyond Benchmarks: Measuring Success for AI Initiatives
Why offline metrics don't predict online success and how to design evaluation systems that align model development with business outcomes.
Salma MayorquinWatch →
⚡Talk · AG2 Community
GitRank: Automating Paper-to-PR with Agentic Code Execution
Technical deep-dive on how GitRank uses AG2 agents to discover, build, and test research implementations from arXiv to live experiments.
Terry RodriguezWatch →
📰Feature · Cerebral Valley
Remyx: Your AI Production Assistant
In-depth interview on the mission behind Remyx, how the founders met at Berkeley, and their vision for the next generation of AI engineering tools.
Salma MayorquinRead →
✍️Blog · Substack
Vibes Don't Scale
Why AI-assisted coding falls apart past the demo stage and what disciplined experimentation looks like when complex code changes are under-determined by simple prompts.
Terry RodriguezRead →

Active in the AI Community Open Source

We contribute open-source tools, datasets, and benchmarks across AI domains and the research community builds on them.

🌐
From landmark paper to open infrastructure
We reproduced DeepMind's SpatialVLM data pipeline entirely with open-source tools, then trained models a fraction of the size, proving small, open models could learn spatial reasoning too. This enabled follow-on NeurIPS publications, new benchmarks, and a wave of open models across the field.
Google DeepMind · CVPR 2024 →
🎓
Datasets adopted by NeurIPS researchers
Oxford and UC Santa Cruz researchers trained and evaluated their models using Remyx open-source datasets.
Oxford / UCSC · NeurIPS 2025 →
🛠️
VQASynth: Open-source data synthesis pipeline
A toolkit for composing multimodal training datasets, used and extended by researchers worldwide to advance model capabilities.
GitHub · remyxai/VQASynth →
🤝
Partner integration with AG2 (formerly AutoGen)
Co-developed the RemyxCodeExecutor for AG2's agentic framework, enabling researchers to explore and test code from any paper.
AG2 Official Docs →
📈
30K+ downloads on community model releases
Lessons learned from publishing open models on Hugging Face and what makes research artifacts actually get adopted by practitioners.
Substack · Myx'd Results →
🌐
From landmark paper to open infrastructure
We reproduced DeepMind's SpatialVLM data pipeline entirely with open-source tools, then trained models a fraction of the size, proving small, open models could learn spatial reasoning too. This enabled follow-on NeurIPS publications, new benchmarks, and a wave of open models across the field.
Google DeepMind · CVPR 2024 →
🎓
Datasets adopted by NeurIPS researchers
Oxford and UC Santa Cruz researchers trained and evaluated their models using Remyx open-source datasets.
Oxford / UCSC · NeurIPS 2025 →
🛠️
VQASynth: Open-source data synthesis pipeline
A toolkit for composing multimodal training datasets, used and extended by researchers worldwide to advance model capabilities.
GitHub · remyxai/VQASynth →
🤝
Partner integration with AG2 (formerly AutoGen)
Co-developed the RemyxCodeExecutor for AG2's agentic framework, enabling researchers to explore and test code from any paper.
AG2 Official Docs →
📈
30K+ downloads on community model releases
Lessons learned from publishing open models on Hugging Face and what makes research artifacts actually get adopted by practitioners.
Substack · Myx'd Results →

Ready to stop guessing?

Start improving your AI system in minutes.

Get started for free
Watch a demo