Case study · ML developer tools · 3 min read
PerceptiLabs: How a new ML tool earned developer trust and built a community
PerceptiLabs™ had a new visual modeling tool and a skeptical audience: code-first ML developers. We built the developer relations and developer marketing strategy with the founders, then ran it: developer experience and technical content, outreach, events including a virtual booth at ODSC™ East 2021, and day-to-day community management that grew a trusted developer community.
Impact at a glance · within one year
2.5x
Community forum members
3.4x
Forum posts
3.1x
Discussion topics, as developers brought more questions and ideas to the forum
1,000+
YouTube® subscribers within a few months
Also earned: A KDnuggets Gold Blog · Features on Kaggle™ and Hugging Face™ · Collaborations with organizations including Red Hat® · A virtual booth at ODSC™ East 2021 that our team helped design
Client
PerceptiLabs, a visual modeling tool for TensorFlow™
Audience
ML engineers, developers and AI practitioners
Engagement
Ongoing fractional developer relations, developer marketing, developer experience, community and event management, and outreach
Led by
Dana Fujikawa, Relaris Founding Technical Lead
The challenge
Earning trust from developers who live in code
PerceptiLabs let ML engineers build TensorFlow models visually, see how each part affected results, and tune models faster than with code alone.
The hurdle: ML developers trust code, and they spot marketing fluff in seconds. A visual tool had to prove itself to them. Three things raised the stakes:
- A small founding team that needed one partner to set strategy and carry it out
- A fast-moving product, with more than 50 public releases in under 18 months, so docs and tutorials had to keep pace
- A crowded field of ML tools competing for the same developers' attention
What we did
One team for strategy and execution
We joined the PerceptiLabs team on a fractional basis and worked in short, flexible cycles, adjusting as the community responded. Our team used the tool hands-on, so every recommendation and every piece of content reflected how developers actually build models.
Strategy
- Personas and messaging for ML developers and engineers
- Website content strategy and technical content planning
- Ad copy, outreach and partnership opportunities
Content and developer experience
- Ongoing technical documentation
- Video tutorials and whitepapers
- Blogs our team wrote for PerceptiLabs, syndicated to popular ML sites
Community
- Community building and day-to-day management
- Forum moderation, answering questions and guiding discussions
- Event planning and management, including helping design PerceptiLabs' virtual booth at ODSC East 2021
- Nurturing early users into active champions
The impact
A community that grew bigger and more engaged
Within a year, the engagement contributed to clear growth in community and product adoption. Membership grew, and members took part more often.
Community growth during our fractional engagement
2x
Likes per post
6.3 → 8.7
Posts per member, on average
Reach and credibility
- 1,000+ YouTube subscribers within a few months
- A KDnuggets Gold Blog, one of the site's most-read posts: PerceptiLabs: A GUI and Visual API for TensorFlow
- Features on Kaggle and Hugging Face, where ML developers already spend their time
- Several important collaborations with organizations including Red Hat
What this means for your team
Reach does not equal adoption.
PerceptiLabs grew because developers found technical content that respected their time and a community where their questions got answered. That is how we work with every client:
- Senior practitioners who use your product hands-on before recommending anything
- Strategy and execution from the same team
- Content, community, events and outreach built to earn developer trust, not just clicks
Are developers adopting your product the way you planned? Let's find out together.
Schedule a free consultation© 2026 Relaris Group LLC. All rights reserved. This case study, including its text, design and analysis, may not be copied, reproduced, distributed or used in any form without prior written permission from Relaris Group LLC. Red Hat is a registered trademark of Red Hat, Inc. YouTube is a registered trademark of Google LLC. PerceptiLabs, TensorFlow, Kaggle, Hugging Face, KDnuggets, ODSC and other company names are trademarks of their respective owners and are used with permission or for identification only.