Inside Anomalo’s Hackathon: What We Built, What We Learned, and How to Run Your Own

From October 21 to October 25, the Anomalo Slack was oddly quiet. As an all-remote team, our Slack is usually buzzing with activity, but for one week, all the engineers were together—in person—for our annual hackathon. This year’s theme? Tooling 🛠

Each year, we pick a new location to meet, and because we’re a team of nature lovers, we decided to take a trip to the vast outdoors. We chose AutoCamp, an Airstream campground in Cape Code, Massachusetts, that sat above the beach, overlooking the water. Each team member stayed in their very own Airstream, and teams used the common tents and campfire pits as breakout rooms, taking breaks walking on the beach or taking the cruisers for a spin.

First lesson? Coding is so much more fun when done outdoors, and morning huddles hit different when you’re sitting around the campfire with a coffee and bagel. 

But that wasn’t the only thing we took away from this trip. Hackathons are a powerful tool not just for getting a ton done in a short period of time but also for resetting team dynamics and strengthening engineering culture. In this post, we’ll share not only what we built but what lessons we flew home with and some tips for organizing a successful hackathon at your company.

Four New Features We Built

Our engineers flew in on Monday, October 21, and flew home on Friday, October 25, giving them only two and a half days to bring their ideas to life and demo them to the rest of the company on Thursday. Despite the incredibly tight timeline, each team was able to develop a complete end-to-end feature – some of which we shipped to production just weeks later!

Feature Flag Service: Taking the Guesswork Out of Flags

Feature flags: powerful but painful. Our old system managed feature flags in the Django admin interface, and while that worked great when we had a small set of customers whose deployments we were all intimately familiar with, as we scaled, feature flags became hard to keep track of, manual, and about as fun as debugging on a Friday night. 

The team brought feature flags into the light and in the App itself. They built a streamlined feature flag service with near-instant updates, bulk changes, and permission controls. Plus, flags now auto-expire — no more zombie flags lurking in the shadows. Managing flags just got fun (or at least less annoying).

Check Recommender: Because Configuring a Check Shouldn’t be Only for Experts

While our automated data quality checks are our bread and butter, we feel strongly that a customer should be able to set up any check they might ever care about on their data. Our product has hundreds of low-code/no-code out-of-the-box checks to help with this, which you can set up in just a couple of clicks. We thought – easy enough, right?

Anomalo is designed for all members of the data team, not just the data experts. In order to democratize access to the data quality checks, the team reimagined the check configuration flow. Configuring a check now requires a single skill: speaking English.

The team built a new RAG/LLM powered check configuration flow that allows you to turn “I want to make sure column X is never empty when column Y is filled in” into a new check in your table. The check recommender even writes SQL for you and executes the full flow to let you confirm you’ve configured everything correctly before saving and shipping the check. Less trial-and-error, more click-and-done.

Exception Dashboard: Debugging Without the Headache

Like most organizations, we have a lot of tools that allow us to catch and track errors. Also, like in most organizations, toggling between Grafana, Sentry, and a dozen tabs to track exceptions was becoming cumbersome. As we started receiving more and more alerts and false positives, it was also becoming harder to separate the signal from the noise. 

This team built a new exception dashboard that consolidates everything into one clean interface. You can see trends over time, read summaries of the trends, and drill down into specific issues all in the dashboard. Allowing engineers to go from “What broke?” to “Fixed it” without breaking a sweat.

Incident Agent (a.k.a. LlamaOps): Your Incident Wingman

Given this year’s theme encompassed both internal and external tooling, there was one feature that became an instant hit with the rest of the engineering team. If you’ve ever been on-call or had to deal with an incident, you already understand why.

Incidents are stressful. Enter LlamaOps, your AI-powered incident buddy. LlamaOps is another LLM-powered feature that fetches data from Notion, Grafana, GitHub, and Slack to help you identify, resolve, and document each incident. It analyzes past incidents and resolutions to help an engineer get to a resolution more efficiently, but it also uses function calling to analyze time series trends and PR merges to identify root cause incidents quickly. 

All of this happens by conversing with LlamaOps right in Slack – where all our conversations already happen.  Bonus: once the incident is over, LlamaOps can also summarize the timeline and all key events and next steps, making retros easier. Less panic, more “handled it.”

Four Takeaways We Flew Home With

As you can see, this year’s hackathon was productive – the check recommender is already out, and over time, we want to ship all four projects to production – but it was also instructive. After Demos on Thursday, our team reflected on what we learned from the time spent together, and four main themes emerged. 

1. Cross-Functional Teams Are the Secret Sauce

In 2.5 days, we built real, shippable features. Why? Because we blended app engineers, ML folks, and platform pros into cross-functional squads. We realized handoffs are productivity killers, and real-time collaboration is rocket fuel. Our plan: more cross-team projects, fewer silos, and tighter feedback loops.

2. Less Slack, More Flow

Fewer interruptions = better code. Shocking, right? During the hackathon, we checked Slack occasionally instead of religiously. We focused, paired, and flowed. Inspired by this, we’re piloting a rotating team liaison role — someone to handle questions so engineers can stay deep in the zone.

3. Developer Experience Isn’t a Luxury — It’s a Necessity

We learned firsthand that improving dev experience pays off immediately. Whether it’s faster CI, better codespaces, or tooling that doesn’t make you want to flip your desk, DX investments help everyone: developers, customers, and the business. Our new mantra: If it hurts, fix it — future-you will thank you.

4. The World has Changed

This year, it felt like each team had a bonus engineer on board: LLMs. They helped us brainstorm, write tests, debug, and clean up code. Two of our projects even used LLMs directly. Refining our approach to AI not only makes us better but also helps us understand how other companies are harnessing LLMs. AI isn’t just a buzzword — it’s a productivity multiplier, and we’re all in.

How to Run Your Own Hackathon: Lessons from the Wild

Finally, we’re feeling so energized by our experience, and we’d love to help you get the same out of our hackathon. Here’s what we learned about making one a success – interestingly enough, most of this applies to making teams work successfully day to day:

1. Pick a Theme That Excites Your Team

A focused theme (like Tooling for us) helps channel creativity and solves real pain points. Whether it’s improving internal systems, exploring new tech, or enhancing customer features, a theme gives your hackathon purpose and direction. Pro tip – pick something that is focused yet could be interpreted in different ways to get the most diverse ideas.

2. Location Matters (Think Outside the Office)

We went off-grid to an Airstream campground. You don’t need tents and campfires (though highly recommended!), but changing the environment sparks fresh thinking and stronger bonds. Whether it’s a retreat center, co-working space, or even a decorated office, a new setting goes a long way.

3. Plan (But Not Too Much)

Pre-vote on Ideas: Let your team submit, vote on ideas, and build teams beforehand so everyone arrives ready to roll.

Set Clear Constraints: We had 2.5 days — tight but doable. Constraints force focus and creativity.

Demo Day is Sacred: Make time to showcase everyone’s work. Demos are fun, build energy, and give well-deserved kudos.

4. Mix Up the Teams

Hackathons are a great excuse to break silos. Pair engineers from different teams — backend, ML, platform — and watch the magic happen. Bonus: cross-functional teams surface hidden talents and build new connections.

5. Minimize Interruptions

Treat your hackathon like a sacred bubble of productivity. Limit meetings, Slack pings, and daily distractions. Need someone to stay on top of things? Assign a liaison to handle questions while the team stays in the zone.

6. Celebrate and Ship

Wrap it up with a celebration, and if possible, ship those projects! Even if the code isn’t perfect, getting ideas into production boosts morale and shows the impact of innovation.

An open runway for innovation

Our hackathon wasn’t just about shipping cool projects (though we did that). It reminded us that investing in tooling and developer joy makes everything better. Happier engineers build better products faster. If that sounds like your kind of vibe, we’re already planning next year’s hackathon.

Come build with us! 🚀