Scaling smarter: Engineering wins at Point in 2024

2024 was a great year for Engineering at Point. We’ve been operating a lean engineering team with a laser focus on delivering high-impact products while maintaining exceptional execution. The results have been outstanding.

The overarching theme of 2024 was growth—our origination volume grew tremendously over the course of the year, while engineering headcount has remained largely the same. This is a testament to our ongoing efforts to raise the bar for teams while maintaining the quality of life that makes Point a great place to work as a software engineer.

This level of growth requires that we major on the majors: delivering exceptional customer experiences that help homeowners understand how a Home Equity Investment (HEI) can improve their financial well-being while continuing to invest in the accuracy and efficiency of our internal processes. 

Behind the scenes, we made one of the biggest technical shifts in Point’s history, moving our infrastructure from Heroku to AWS. This shift was motivated by our commitment to security and allows us to take advantage of AWS’s fine-grained control of network access and other security features. As part of this shift, we’ve adopted Infrastructure as Code (IaC), meaning that all infrastructure changes are peer-reviewed and auditable. AWS also provides an incredibly rich set of products that expands our architectural design options and empowers faster innovation.

Major focuses and wins

We accomplished far too much in 2024 to cover in a single blog post, but below are eight major themes—and wins—that we focused on over the year.

Improving customer experiences through technology

From interactive tools to smarter document handling, the team has prioritized customer empowerment and self-service. 

At Point, we care deeply about our homeowners and are committed to helping them make informed decisions. Since Home Equity Investments (HEIs) represent a new class of financial product and a fresh way to think about the value of home equity, we’re investing heavily in dashboards and interactive tools. These features help homeowners understand how this product can impact their finances and guide them seamlessly through the application process.

Impact: Improved customer satisfaction, reduced support queries, and increased transparency in user interactions.

Key projects:

• New homeowner dashboard release
• Interactive payoff estimator
• Automated document validation
• AI-powered document verification system

What's a behind-the-scenes moment that perfectly sums up how this project came together?

“A behind-the-scenes moment that perfectly captures how AI document verification came together was an impromptu whiteboarding session during a company onsite. Tech leads and product managers from multiple teams came together to lay the foundation for the design. From there, the team embraced a scrappy, fast-paced approach, prioritizing customer experience and continuous iteration over perfection.”

‍Krishna Abburi, Senior Manager, Software Engineering

Accelerating decision-making with AI and automation

Leveraging AI and automation, the team has transformed complex decision workflows, enabling near-instant processing and precision. 

Since HEIs have no monthly payments, Point can work with homeowners with more diverse financial backgrounds than traditional mortgages. Consequently, ensuring that these homeowners are well-qualified for an HEI is more complex. We are investing heavily in automation to remove manual reviews where possible, increasing the consistency of our underwriting processes and reducing the time and effort required to review homeowners’ applications.

Impact: Faster application processing, reduced manual effort, and improved accuracy.

Key projects:

• Automated docket qualification system
• Advanced risk detection system
• Enhanced credit risk management system
• Automated "marked ready" process

What was the trickiest problem you solved, and how did you crack it?

“We had an issue early on in the development of the decisioning task framework in UW that was seemingly inexplicable. Sometimes, required tasks would be created, and other times, they would not. I had a pair-debugging session with Chris George, where we worked step by step through the code, executing line by line. We only confused ourselves more because the behavior was different every time. Eventually, we found the culprit—an @instance_variable within a class << self block. It turns out this is equivalent to a @@class_variable, which means that the decision framework would store the information for one task, and then it would leak into the next. This perfectly explained what had been seemingly inexplicable. And, of course, as is common in these cases, the intuition for what the problem might be came just around the time we were getting ready to give up for the day.”

Ryan Minniear, Technical Engineering Lead

Transforming portfolio management

Unified monitoring platforms and automated risk detection have elevated portfolio visibility and operational efficiency. 

With a large and growing portfolio of HEIs, it’s vital that we actively monitor our assets for tax delinquencies, expired insurance, and defaults on senior liens. In 2024, we significantly improved our portfolio monitoring capabilities, integrating data from many different sources and building tools that allow our servicing agents to quickly investigate and take action on potential defaults to minimize the risk of loss.

Impact: Reduced risk exposure, improved investor confidence, and enhanced team productivity.

Key projects:

• Service monitoring platform
• Property monitoring & default triage
• Subservicer integrations

Streamlining underwriting workflows

Modernized underwriting systems have streamlined processes, integrated advanced validation, and enhanced data handling. 

Where we can’t automate, Point engineers work closely with our underwriting counterparts to identify process improvements and build solutions that enable them to work more efficiently. In 2024, we removed bottlenecks in supporting debt payoffs, reviewing title findings, ensuring the accuracy of appraised home values, and more. 

Impact: Improved accuracy, reduced manual review time, and enhanced workflow efficiency.

Key projects:

• Modernized debt payoff system
• UW 2.0 taskifying title and mortgage facts
• Automated occupancy detection
• Appraisal process optimization

Who on the team had a game-changing idea, and what was it?

“Glen, our Lead Product Manager, realized that our docket ops workflow follows a pattern: “What decision needs to be made next?” -> “What info do we need?” -> “How do we get the info?” -> “Do the tasks to get info” -> “Make decision” -> repeat. This understanding revealed two key properties: data and tasks must be prioritized, and decisions don’t care about tasks but only available data.”

Medei Kitagaki, Staff Software Engineer

Strengthening data integrity and analytics

Enhanced data architectures and integrated tools ensure precise reporting and actionable insights. 

At a startup, things change fast, and systems often grow organically to support new requirements and tight deadlines. This year we invested time in several product-facing projects to strengthen several of our core data models—incorporating a more extensive understanding of our business needs than when these data models were originally developed. 

The result? A data layer that is easier for everyone to understand. This promotes a culture of data-driven decision-making and empowers people throughout the company to make use of self-serve data analytics tooling.

Impact: Reduced data inconsistencies, faster analysis, and improved reporting capabilities.

Key projects:

• Enhanced credit risk management system
• Subservicer integrations
• Rearchitected bulk tasks system

Innovating infrastructure to scale

Migration to AWS and adoption of Infrastructure as Code have laid the foundation for robust, scalable systems. 

As highlighted in the introduction, we shifted our infrastructure from Heroku to AWS during the first half of the year. In addition to giving us more fine-grained control over our security posture and better options to scale our production workloads, this effort paid off almost immediately as we were able to design our Property Monitoring system around a cloud-native, serverless architecture. This enabled us to reduce the estimated delivery time by 50% with a more fault-tolerant solution than we could have obtained before.

Impact: Improved security, reduced deployment times, and enhanced scalability for future growth.

Key projects:

• AWS migration
• Property monitoring & default triage

Pioneering next-generation testing strategies

Through shift-left testing, automation, and innovative practices, the Quality Engineering team has accelerated feature delivery without compromising quality. 

We have retrained and upskilled our primarily manual QA team to focus on building test automation frameworks in Playwright that can be executed within a CI/CD pipeline based on GitHub Actions. As a result, we have been able to shift from releasing twice a month to on-demand multiple times a week. The outcome? Happier developers who ship code faster with fewer bugs.

Impact: Faster feedback cycles, reduced bugs in production, and higher-quality features ready for release.

Key projects:

• Shift-left testing strategy
• Daily & pre-merge regression tests
• Test performance improvements
• Feature flags testing

What’s one accomplishment that makes you especially proud?

“I'm very proud of our progress in adopting trunk-based development. Thanks to automated tests running at the pre-merge and merge queue levels, the time required to certify RC has significantly reduced. Now that the Service team operates under at-will deployment, a shorter RC certification time has become invaluable.”

Jasmine Wright, Senior Software Development Engineer in Test

Simplifying complexity for seamless servicing

Consolidated tools and workflows have simplified servicing processes and improved end-to-end visibility. 

Leading into 2024, most of our servicing workflows were performed manually, and information was recorded in our servicing system after the fact. During this past year, we have invested heavily in a platform that can grow to support the wide variety of processes managed by our servicing team. 

The core of this framework is our Task & Workflow system, which allows us to define business processes in terms of specific tasks and sub-tasks, assign an owner to each task, and track the results. These tasks are then composed into workflows—for example, to support homeowners impacted by natural disasters like wildfires. The result of this system is more efficient work, fewer errors, and at-a-glance visibility of what the entire team is working on. 

Impact: Increased team efficiency, reduced manual tracking, and better compliance.

Key projects:

• Service monitoring platform
• Default workflow system launch
• Subservicer integrations

Looking ahead

As you can tell, 2024 has been amazing. The scaled delivery and execution we have achieved is the result of an extraordinary team of software engineers, along with a strong leadership team consisting of Bavyaa Vasudevan, Eric Pierce, Kiran Deep, Krishna Abburi, Michael Smith, and Scott Moore. Their vision, mentorship, and dedication have been instrumental in driving our major technical initiatives and fostering a culture of innovation and excellence.

2025 is shaping up to be even better. 

This year, we continue to build on the foundations we’ve laid—engaging homeowner experiences, outstanding operational efficiency, and best-in-class servicing operations—while doubling down on our investments in security and technical excellence. 

One exciting addition to our roadmap in 2025 is research and development focused on the adoption of AI. While highly regulated industries like real estate and financial technology present challenges in adopting AI, we are convinced that this will be a transformative technology. We began investing in AI to support Intelligent Document Processing in 2024. Handling printed documents is one of the highest friction points in real estate transactions and contributes significantly to the cost of these transactions for homeowners. 

Beyond document processing, we are exploring opportunities to create more engaging self-service opportunities for homeowners and reducing the tedious, error-prone manual work that is endemic within the home finance industry. Our research this year will change the trajectory of the HEI industry for years to come.

As Point continues to scale in 2025, so will our Engineering team. If you want to join us in making homeownership more valuable and accessible, keep an eye on our careers page and follow us on LinkedIn.