Jake Gribschaw didn’t set out to become a software engineer. He fell for coding almost by accident after majoring in chemistry, by way of networking and system administration courses, a few computer science classes, and ultimately, the rush of solving logic puzzles. “Solving problems in code was just really gratifying,” he says. “In my last year of college, I restarted the computer science club. We’d talk about the industry and help each other out. That’s when I knew I really loved this work.”
Frustrated with the prospect of extending his degree by three years to formally complete a CS major, Jake pivoted. He joined Hack Reactor’s software engineering bootcamp during what he calls “the golden age” of immersive programs, spending 16 hours a day, six days a week for several months to accelerate his path to becoming a software engineer.
That intense sprint paid off: His first role in tech was at Apple and since then, he’s charted a winding, fulfilling course through software engineering, data, testing, and leadership roles — including his current position at Chime, where he’s building AI tools to improve developer velocity.
Engineering at Every Altitude
After building a strong foundation in software engineering, Jake was drawn to leadership — first as a team lead, then as an engineering manager. The transition gave him a broader view of the business and a new appreciation for how engineering work fits into the bigger picture.
“Taking on a leadership role really opened my eyes,” he says. “I used to think, ‘Hey, I’m building a thing, and that’s awesome!’ But when you manage, you see how everything connects. You get why documentation matters. You understand how to advocate for your team and align with other parts of the company.”
Still, after a couple of years in management, Jake made the deliberate decision to return to an individual contributor role. He missed the hands-on problem-solving and faster feedback loops that come with engineering work.
“Management has these feast-and-famine cycles — there are times when there is so much to do and very little time to do it, and you might need to spend many extra hours delivering for your team,” he says. “And then there are times where there is very little external pressure, and you learn to take that time to plan and think forward. You also tend to get very large wins but far between. As an IC, your progress is more consistent, and if you fall behind, you can put in the time and push through. I missed that rhythm.”
Now back in the code, Jake continues to bring a systems-level mindset to his work — whether he’s designing tools or mentoring teammates. And while he’s currently focused on building developer-facing AI products, his passion for learning never stops.
“I’m always looking for what’s next,” he says. “That’s what keeps it fun.”
Building Chime’s High-Quality Culture
When Jake joined Chime®, he was the very first software engineer in the test organization. “There wasn’t an engineering services team yet,” he explains. “Just a few traditional QA folks and a big opportunity to define what quality engineering could look like here.”
Jake got to work, building a team, hiring, and moving into a management role for a group of contractors who required strong technical leadership. “That was a learning experience,” he recalls. “Leading a group of engineers who aren’t looking for you to jump in and do the work yourself — but instead to set direction, build systems, and create clarity.”
Over time, Jake helped evolve the group from a small collection of testers to a centralized team that supports Chime engineers across the org. “Whether we’ve been called quality services, testing services, or quality engineering, our mission has always been the same: help Chimers deliver things more safely and quickly.”
That foundational focus on safety, stability, and security hasn’t changed — even as the org around it has evolved.
AI As an Accelerant: the Story of Beacon
At the start of 2025, Jake heard the call from Chime engineering leadership to explore AI tooling. And he knew exactly where to focus: code reviews.
“We saw around 2,000 code reviews a week,” he explains. “That’s a huge time sink — one engineer writes a PR, another stops what they’re doing to review it. If we could reduce even a fraction of that overhead, we’d get massive efficiency gains.”
So Jake began building an AI-powered code reviewer. First he tested it on himself. Then, slowly, he expanded it to his team. The tool — now known as Beacon — was live and in use across Chime in just eight weeks.
“Beacon is deployed in GitHub. Any Chimer can add a beacon/review label to their PR,” Jake explains. “It analyzes the code, applies some rules and style checks, and returns helpful, consistent feedback.”
The development process was fast because Jake knew the space well — and because he used AI tools like ChatGPT, Virtual Studio Code, and GitHub Copilot to help write Beacon itself. “When you’re passionate about something, and the tools are this good, you can move quickly.”
Key to Beacon’s success was Jake’s philosophy around AI: It should be a thoughtful tool, not just a magic button. “You need to break the problem down into discrete tasks the AI can actually perform well,” he says. “That’s what makes it effective.”
Secure, Smart, and Cost-Aware
When building Beacon, Jake approached the project with more than just engineering precision — he brought a leadership-informed lens to balancing tradeoffs. “You’re not just asking, ‘Can I build this?’” he says. “You’re thinking about privacy, cost, scale, and impact.”
With Beacon, those questions took concrete shape. First came security: Because the tool interacts with source code — not PII — it avoids the most sensitive data concerns. But Jake still had to ensure Beacon could operate safely within Chime’s infrastructure. Thanks to existing contracts with AWS and the company’s GenAI Gateway, robust privacy protections and redaction tools were already built in. “It was a great accelerant to have the security guardrails in place,” he says.
Next came cost: Jake designed Beacon’s rollout to be gradual and opt-in, starting with his own pull requests. “We did a slow rollout, so we could track spend as usage grew,” he explains. “Each call might cost a cent or two, but developer time is expensive. If we can save time and prevent even one bug from reaching production, that’s worth it.”
To build awareness around Beacon and Chime’s growing AI tools, Jake developed presentations at both company-wide and conference scales, where he fielded questions from technical audiences. “I enjoyed presenting. It’s gratifying to have a room full of smart people interested in something you’ve created.”
Finally, there was impact: Beacon wasn’t just built with AI — it was built to scale with AI. Jake prioritized metrics and observability from the start, helping the team evaluate real-world effectiveness and inform ongoing tuning.
Today, a significant percentage of all Chime pull requests use Beacon and adoption is growing. Chime recently recognized Beacon’s impact and presented Jake with the Apollo Award for Innovation in AI. “It’s been exciting to build something that’s not just technically interesting, but actually making life better for other engineers,” he says. “That’s where it all comes together.”
What’s Next?
Looking ahead, Jake is excited to continue growing AI adoption through custom agents and bespoke platforms available to developers throughout Chime. These custom built AI tools have already reduced the operational burden on engineers.
“As a manager, I sharpened my ability to identify roadblocks and friction points in the development life cycle. Now as an engineer, these skills allow me to identify areas where AI can be implemented most effectively to reduce the burden on my peers, freeing them to focus on the higher levels of design and implementation, while acting as a force multiplier.”
“I’m a big believer in keeping things simple. With the right tools and the right people, you can build powerful, efficient systems without overcomplicating them.”