Platform
Resources
Hackathons are a Statsig tradition.
At Statsig, Hackathons are quarterly events where employees set aside their usual work to build anything they want—whether it’s an experimental feature, a wild automation, or a just-for-fun internal tool.
Over the course of two days, engineers, designers, marketers, and PMs team up, build fast, and demo their projects live. Sometimes these hacks make it to production. Sometimes they just make us laugh. Either way, it’s a core part of our culture.
Let’s take a look back at some of the highlights from our Q1 2025 Hackathon.
Note: Unless otherwise stated, these are prototypes or internal tools. If something catches your eye and you’d like us to prioritize it—let us know!
Skip ahead:
🧠 AI and automation
Projects that leverage AI or automation to make using Statsig easier, smarter, or more powerful.
Statsig MCP server
Presenters: Brock Lumbard, Jairo Garciga, Andrew Lee, Alisha Lakkad, Oliver Zheng
Summary: Think of this as an AI assistant with console-level access. The MCP Server watches your projects and summarizes what it sees—then lets you ask it for deeper insight, including summaries, metrics, and suggested actions.
In one example, the team asked it to locate stale gates and remove them. The tool found the gates, dug into the code, and took care of the cleanup—all via the console API (no extra SDKs required).
Access is controlled via your API key, so actions can be scoped and limited as needed. Better yet, the tool will be open source—anyone will be able to install and run it themselves.
Statsig setup wizard
Presenters: Andre Terron, Lior Barak, Daniel West, Bobby McDonald
Summary: Installing Statsig just got way easier.
This project introduces an AI-powered setup wizard that helps users integrate Statsig into their projects with minimal effort.
In the demo, the team used a fictional “spellbook” app. With a single npx command, the setup wizard asks for your Statsig console key and OpenAI key, installs the necessary packages, and then uses AI to read and edit your codebase—automatically injecting Statsig providers and placing keys where they belong.
In Cursor, Andre showed how the tool updated the project and immediately started sending events into Statsig. This will make Product-Led Growth (PLG) onboarding dramatically smoother.
Sidecar surveys
Presenters: Pablo Beltran, Tyler VanHaren
Summary: Pablo and TVH are making Sidecar even more usable—especially for marketers.
Their Sidecar extension allows users to visually modify a survey and click a button to “match style of website,” which automatically applies the CSS styling. There’s even a session recording of someone filling out the survey in real time.
The goal? Make Sidecar a truly no-code experience. You can launch and customize experiments—complete with visual styling—without touching any code. And yes, all events still show up in Statsig.
Related: Announcing Statsig Sidecar: Website A/B tests made easy
SDK autocomplete
Presenter: Andre Terron
Summary: Andre’s hack makes working with Statsig SDKs a lot smoother.
If you forget the name of a gate while coding, you typically have to jump back to the console. Andre prototyped a tool that uses our CLI to generate TypeScript definitions for all your gates. Once generated, your IDE can autocomplete them and even warn you if a gate is disabled.
It’s just a proof of concept for now, but it could save a lot of context-switching.
A/B prompt testing
Presenter: Joe Zeng
Summary: Joe’s hack brings A/B testing to the world of LLMs—making Statsig a true end-to-end platform for model and prompt evaluation.
He built new entities inside Statsig for prompts, test cases, and models. For the demo, he created a Statbot prompt, defined what good responses should look like, and ran offline evaluations using Statsig’s Eval API.
The test runs output pass/fail results directly inside the console. In the future, Joe plans to tie in real user events to help automatically retrain and optimize prompts over time.
Related: Experimenting with generative AI apps
Autotune statscript jobs
Presenter: Brent Echols
Summary: Brent created a custom query language to analyze Dagster jobs and their memory usage. It pulls data about job performance and writes it to a file—potentially helping with tuning and optimization.
📈 Analytics and experimentation
Tools that expand Statsig’s experimentation and data analysis capabilities, or improve how results are presented and explored.
Quasi analysis
Presenters: Shubham Singhal, Michael Makris, Nicole Smith, Cynthia Xin
Summary: Most feature flags are used as simple toggles, which makes classic A/B testing tricky—especially when something’s rolled out 100%. This project introduces a way to evaluate those rollouts using quasi-experimental analysis.
Nicole recently enabled user journeys for all WHN customers. In the demo, Shubham selected a metric and compared its performance before and after the rollout. This allows teams to answer questions like, “Did user journeys actually move the needle?”—even after a full rollout.
Write to WHN
Presenters: Liz Obermaier, Nicole Smith
Summary: This tool makes funnel analysis in your warehouse a lot more actionable.
Instead of exporting CSVs to analyze drop-offs, conversions, or funnel stages, you can now create and query personal tables directly in your data warehouse.
Each user gets their own private set of tables to avoid cluttering shared projects, and all tables are searchable and previewable in-app.
Histogram view for unit-level metrics
Presenter: Craig Sexauer
Summary: Craig introduced a new visualization for unit-level metrics: histograms.
While line charts and tables tend to get the spotlight, histograms are a favorite among data scientists, especially when it comes to understanding distribution and variance. Several users have requested this in the past—now it exists as a hackathon prototype.
Charty
Presenter: Craig Sexauer
Summary: Charty is a game that tests your chart-reading skills—and reveals some surprising insights in the process.
Players are shown a series of visualizations (bar, pie, donut) and asked to guess the correct percentages. Their answers are timed and scored for accuracy.
Turns out:
Bar charts produced nearly 2x the user errors compared to pie charts
Users spent 50% more time guessing bar charts
Even donut charts had better performance than bar charts
Fun and informative.
Related: Play Charty it's really fun
Statly
Presenter: Yuzheng Sun (YZ)
Summary: What if running experiments didn’t require reading any docs?
YZ’s Statly is a playful but practical experiment in simplifying A/B testing for early-stage startups. The idea: strip everything down to the essentials, offer simple experimentation tools, and then hand off more advanced workflows to Statsig Lite.
It’s mostly a marketing thought experiment—but one that could help expand experimentation to folks who don’t have time to get technical.
Statly ultimately funnels users to Statsig Lite.
Tabs in Metrics Explorer
Presenters: Alex Coleman, Daniel Kim
Summary: Metrics Explorer just got tabbed.
Instead of opening new browser windows for each metric, you can now open and organize multiple tabs within Metrics Explorer. You can name your explorations, and each tab inherits the name to make it easy to jump between views. Just like a browser—only purpose-built for digging into data.
🛠️ Infra and developer experience
Infra upgrades, dev tooling, and performance optimizations that improve backend efficiency or developer workflows.
Oncall improvements
Presenters: Pierre Estephan, Nick Jiang
Summary: The oncall page needed a glow-up, and Pierre and Nick delivered.
Previously, oncalls were configured via dynamic configs and were mostly limited to assigning a person and a time. The new interface is a full-featured scheduling dashboard:
See current oncalls and upcoming shifts
Ping engineers from within Statsig
Edit and override shifts with drag-and-drop
Access the oncall API to explore related entities and schedules
View a preview of your changes before committing
Big upgrade for oncall quality-of-life.
Traces
Presenters: Scott Richardson, Jason Wang
Summary: Statsig now supports trace visualization using OpenTelemetry.
Jason and Scott added a way to customize spans generated by OTel, which means we can now see them clearly in a unified interface. The trace viewer shows all requests flowing through our system as a clickable waterfall, with span metadata and JSON previews for each step.
Previously, it was tough to correlate backend events or trace requests from start to finish.
HostMap
Presenter: Solomon Choe
Summary: Solomon built a visual representation of infrastructure—both ours and potentially customers’—called Hostmap.
Each node represents a container, and node colors are mapped to metrics of interest. Inspired by monitoring tools like Dynatrace and Datadog, Hostmap uses WebGL for high-performance rendering and aims to help teams quickly spot inefficiencies, monitor activity, or identify cost-saving opportunities.
Figma icons library
Presenter: Brian Do
Summary: Brian streamlined how icons make their way from Figma into the Statsig codebase.
Before: Icon imports were manual and prone to inconsistencies.
Now: There’s a standardized frame in Figma and a script developers can run to pull everything in automatically—no naming mismatches, no guesswork. Just clean imports, every time.
Docs revamp
Presenter: Jina Yoon
Summary: Jina kicked off a revamp of our documentation experience after noticing a few common issues: outdated screenshots, unclear navigation, and general confusion about where new content should go.
Her plan includes a more structured layout, better navigation, and clearer formatting using tabs—making docs more usable for both readers and writers.
Stay tuned for cool updates to our documentation!
Status page revamp
Presenter: Karan Luthra
Summary: Karan redesigned our status page to be far more powerful and integrated.
The new version runs live health checks against our APIs, includes an RSS feed, and can send real-time updates directly into Slack.
You can also create incidents from one central app—with templates for writing public updates, selecting impacted components, and broadcasting alerts across Slack and the status page (complete with a banner).
💫 Culture and brand
Fun or internal-facing projects that focus on company culture, visual identity, or team-wide quality of life.
Pop-up thrift shop
Presenter: Sami Springman
Summary: Sami turned the office into a temporary thrift store. Employees could browse and take home clothes from the pop-up until Friday—after which any remaining items would be donated. A feel-good mix of sustainability and style.
Collective noun project
Presenter: Julie Leary
Summary: What do you call someone who works at Statsig?
Julie crowd-sourced ideas from the team, filtered out any difficult ones, and narrowed it down to a top 5. After a company-wide vote, a winner emerged:
Statizens!
Long live the Statizens. Imagine being called "Siggies." Yikes.
Stat snack tracker
Presenter: Alina Zhu
Summary: Snacks matter. Alina built a tracker that scrapes our food provider’s website to see what snacks are currently available.
It doesn’t stop there—she also built a churn optimizer and created a snack ordering “alphabet” that separates must-haves from nice-to-haves. Finally, some snack science. 🧪
Reddit ads exploration
Presenters: Gina Song, Jina Yoon
Summary: Reddit ads are notoriously hard to get right—but the two G(J)inas are on it.
They explored a range of Reddit ad styles:
The overly polished (and ineffective)
The meme-y
The ones that try to blend in like real user posts
And the rare gems that interact authentically with the subreddit
While no ads have launched yet, the project aims to craft a tone that feels native to Reddit—without being cringey or salesy.
Staticons open source
Presenters: Jessie Ong, Kenny Yi
Summary: Statsig’s icon set—Staticons—is now open source! You can browse the library at statsig.com/staticons or grab the React package on GitHub to start using them in your own projects.
Go forth and icon. 🧑🎨
Related: Introducing Staticons
Poster room redo
Presenters: Cat Lee, Jessie Ong
Summary: The posters in our office were falling off—and with our updated brand, it was time for a refresh anyway.
Cat and Jessie gave the poster room a full revamp, replacing old prints with new designs that reflect Statsig’s current look and feel. A clean upgrade for the vibes.
That’s a wrap
We hope this gave you a fun look behind the curtain at what our Hackathons are really like. If you’ve ever wondered where certain features or quirky touches come from—chances are, they started as a hack.
Some of these projects may ship. Some may never leave the prototype stage. But all of them represent the creativity, curiosity, and chaos that make Statsig… Statsig.
Thanks for reading!
Request a demo
