AI is doing some remarkable things.
Historians used AI to read illegible ancient scrolls that survived the eruption of Mount Vesuvius in A.D. 79.
Google’s DeepMind AlfaFold 3 allows scientists working in fields like vaccine development and plastic recycling to assess the complex structure of proteins — a process that used to take years.
However, AI is a double-edged sword — especially generative AI.
Training deep learning models is known to have a hefty environmental footprint, and a survey from Resume Now found that 90 percent of young workers fear AI burnout. The duality of AI presents engineers with a challenge: What is the best way to add the power of AI into their products while staying focused on the improvement of people’s lives?
Built In spoke with four engineers who had some ideas.
“Many of our customers have shared how the recent AI hype has resulted in a flood of so-called ‘time-saving tools’ that, in reality, are nothing more than basic email suggestion generators,” said Matt Lowe, software engineer at EliseAI. “Understandably, this has led to AI fatigue amongst our customers.”
Lowe and colleagues are building a CRM system on top of an effective generative AI that the team already has up and running.
“It’s incredibly rewarding to demonstrate firsthand how our CRM truly saves time — by surfacing only the tasks that need attention, optimizing tour schedules and enabling bulk actions for common, repetitive workflows,” Lowe added.
Meanwhile, Ongun Savas, a front-end software engineer with Constrafor, is using AI to help construction subcontractors in their bidding process.
“It’s a game-changer,” said Savas. “Building tools that directly improve the financial health and operational effectiveness of construction businesses and being part of such a big impact project is highly motivating for me.”
Lowe and Savas, along with two other local engineers, shared with Built In the details of what they are working on and how each project is bringing out the best in AI.
Constrafor is a SaaS and fintech innovator built for construction.
What project are you most excited to work on in 2025?
I’m most excited to contribute to the development of the subcontractor financial tooling. Currently I am working on the lien waiver collection feature and this will lay the foundations of the initiative we are trying to achieve in 2025. I am also looking forward to working on the integration of AI for bidding support. What’s particularly exciting is the potential impact on our clients. Lien waivers are a critical but often cumbersome process in construction. Streamlining this through intuitive front-end design will alleviate a significant pain point for subcontractors.
Furthermore, leveraging AI to empower subcontractors in their bidding process is a game-changer. Building tools that directly improve the financial health and operational effectiveness of construction businesses and being part of such a big impact project is highly motivating for me.
What does the roadmap for this project look like?
The roadmap involves iterative development. I’m currently focused on delivering the lien waiver collection feature as the initial test, followed by other features that our clients need regarding construction finance. I am particularly excited about the integration of AI-driven bidding support. We’ll prioritize user feedback and data analysis to ensure they meet the needs of our clients.
I’ll be working closely with back-end engineers to establish seamless API integration. I’ll partner with product managers to define feature requirements, and with designers to create a user-friendly interface.
“I am particularly excited about the integration of AI-driven bidding support. We’ll prioritize user feedback and data analysis to ensure they meet the needs of our clients.”
Potential challenges include ensuring accuracy for the AI models and managing the integration of various data sources. To overcome potential challenges, we’ll implement robust data validation and testing procedures. We’ll work with an agile development methodology by communicating frequently to address integration challenges proactively.
What in your past projects, education or work history best prepared you to tackle this project? What do you hope to learn from this work to apply in the future?
My experience as a software engineer at Google provided me with a solid foundation in front-end development using React, Angular and Typescript. I’ve worked with technologies that are transferable to my current role at Constrafor, and adopted industry best practices that will help me thrive in my career.
Specifically, my work on the Copilot marketplace for Waze prepared me to handle the intricacies of building robust and scalable front-end applications. My experience working in agile environments has sharpened my collaboration skills and ability to adapt to evolving project requirements.
I hope to deepen my understanding of our in-house web components, Angular and especially AI integration in user interfaces through this project. This knowledge will be invaluable in my future work, particularly as AI becomes more prevalent in software development.
Ro is a direct-to-patient healthcare company with a mission of helping patients achieve their health goals by delivering the easiest, most effective care possible.
What project are you most excited to work on in 2025?
Our team is working on an exciting LLM-based pipeline to triage the millions of incoming patient messages we receive each year. The goal is to help our incredible team of nurses, providers and patient advocates by intelligently routing messages to the person most qualified to help with that specific question or need. This will give our care teams more leverage and save them time while supporting our patients. This is compelling because it directly solves an immediate, real-world problem.
What does the roadmap for this project look like?
Our goal is to streamline response times for patients while reducing manual triage time for our care teams. We’ll start with classifying and routing patient messages to the correct team of responders. From there, we will develop a series of LLM prompts designed to identify situations and route them with high accuracy to specialized teams. We’ll also focus on using LLM to enable context gathering, to pull in relevant details to summarize the issue concisely for the responder. This will reduce time spent searching for information and allow for quicker, more informed decision-making.
“Our goal is to streamline response times for patients while reducing manual triage time for our care teams.”
To bring this vision to life, we will collaborate with multiple teams across the organization — from back-end and front-end engineers to designers to operations specialists to data analysts. This highly cross functional group will play key roles to ensure this new capability provides a seamless user experience. We will continuously review patient messages with human experts to refine the model’s accuracy. Coordinating these audits is a critical part of the feedback loop that will inform our iterative improvements to deliver a better experience for both patients and our care team members.
What in your past projects, education or work history best prepares you to tackle this project? What do you hope to learn from this work to apply in the future?
Over the past decade, I’ve worked on a wide range of machine learning projects, spanning natural language processing and computer vision. However, my deepest focus has always been on NLP — originally working with small language models and continuously pushing the limits of what was possible.
I identify strongly as a prompt engineer, having engaged in some form of prompt engineering as early as 2016, long before the term was widely recognized. My experience working with early deep learning architectures gave me a strong foundation in understanding model behavior, and an intuition for crafting the inputs and harnessing the outputs of models to coax out new product possibilities. These skills have only become more relevant and valuable today with the rise of powerful transformer-based architectures.
Through this project, I hope to refine my ability to bridge the gap between AI capabilities and real-world operational needs. Ultimately, this work will shape how I approach AI-human collaboration, not just in healthcare but in any domain where AI needs to function as an assistant rather than a replacement.
January is a fintech company that sets a new standard for humanized debt collection. Its tech-enabled platform improves recovery rates and sets creditors and borrowers up for success.
What project are you most excited to work on in 2025?
I’m excited to continue leveraging improvements in LLMs to help our borrowers. I know, everyone says that, but we’re applying it in ways that directly improve financial stability for millions.
One of my team’s responsibilities is improving our contact center’s tooling — so when we make our agents’ jobs easier, we’re also helping consumers get better assistance. Every week, we spend time shadowing our agents, getting visibility into the challenges they’re facing and seeing how our improvements help them. The tight feedback loop helps keep us focused on what matters.
In 2025 we have exciting plans to surface smarter, context-aware insights to our agents, helping them guide consumers through their financial journeys. A few weeks ago I worked on a hackathon project that used an LLM to generate relevant insights, and present it to the agents at the moment they need it. But that was just a demo. Next quarter, we’ll take that foundation and build something production-grade, with the right guardrails and oversight.
What does the roadmap for this project look like? What challenges or blocks do you anticipate? How do you envision overcoming those challenges?
We are still in the brainstorming phase of this project, so it isn’t fully spec’ed out yet. To make a more robust, production ready version of our hackathon project, we’ll need to work with our talented analytics team. They plan to use an ML model to add context to consumers’ profiles. That foundation will help us prioritize the most relevant insights for agents so they aren’t overwhelmed with noise.
We may leverage this with a CDP and telephony integration to give agents actionable information in real time. The goal isn’t just to solve consumers’ current issues, but to proactively equip agents with context that helps borrowers navigate their debts more effectively.
One technical challenge will be scaling efficiently — January works with millions of consumers, so we need to make sure we surface insights where they have the most impact. A broader challenge is driving adoption among agents. We don’t want to disrupt their workflows or slow them down. To get this right, we’ll work closely with agents from day one and iterate alongside them before rolling it out more broadly.
What in your past projects, education or work history best prepares you to tackle this project? What do you hope to learn from this work to apply in the future?
Last quarter, I led a project that was probably the most fulfilling technical work I’ve ever done — an AI-powered tool that assists consumers with January’s application. That work built foundational AI infrastructure, which we’re now using to accelerate future projects. We set up systems and services for regression testing, guardrails against prompt injections, monitoring quality in production and gathering insights from consumer interactions. The investment lets us move faster and more safely when deploying AI-driven improvements.
“Last quarter, I led a project that was probably the most fulfilling technical work I’ve ever done — an AI-powered tool that assists consumers with January’s application.”
My team worked so well together — we collaborated with design and various stakeholders, aligned on a plan, split up the work, and shipped a tool that’s now helping consumers every day. All in less than a quarter! Every week we review how it’s being used, and it’s awesome to see our work making a difference. I’m excited to tackle even more ambitious projects this year, building AI-assisted tooling that helps us execute on our mission of helping consumers achieve financial stability.
EliseAI is a conversational AI startup based in New York City, using machine learning to automate business conversations.
What project are you most excited to work on in 2025?
I’m most excited to continue building out our centralized CRM platform with AI integrations that genuinely help agents save time on repetitive, manual tasks, allowing them to focus more on operations that drive community success. EliseAI is in an exceptionally rare position where we’ve spent years developing effective and practical generative AI, and now we face the much easier task of building a CRM on top of those tools, rather than the other way around.
Many of our customers have shared how the recent AI hype has resulted in a flood of so-called “time-saving tools” that, in reality, are nothing more than basic email suggestion generators. Understandably, this has led to AI fatigue amongst our customers. That’s why it’s incredibly rewarding to demonstrate first-hand how our CRM truly saves time — by surfacing only the tasks that need attention, optimizing tour schedules and enabling bulk actions for common, repetitive workflows.
What does the roadmap for this project look like?
Currently, I see this project being a major part of my day-to-day over the next three months, given our current goals and commitments to our partners. One of the most enjoyable aspects of this project is the high level of collaboration as it involves working with nearly every department at EliseAI, including design, operations, strategy and product while also iterating with external developer teams for integrations and gathering customer feedback on a weekly basis.
At this point, we’ve addressed every foreseeable roadblock. However, with any project of this scale, there will be surprises along the way. More than anything, I’m excited to tackle those challenges because so many people are invested in this project. I have full confidence that our team will collaborate to develop innovative solutions that continue advancing technology in the housing industry.
"I have full confidence that our team will collaborate to develop innovative solutions that continue advancing technology in the housing industry."
What in your past projects, education or work history best prepares you to tackle this project? What do you hope to learn from this work to apply in the future?
Over the past year at EliseAI, the most valuable preparation I’ve had has come from two key factors: the ability to easily engage with our customers and a strong focus on developer velocity. From my very first month on the team, I was on calls with customers at every level of the industry and meeting with clients in person across the country. This experience has given me a much deeper understanding of which problems need to be solved and why they are important.
In my experience, there’s only so much context that can be gained from reading a lengthy list of product requirements. However, being able to ask real-time questions to the people who use the platform daily and understanding what is and isn’t important to them significantly improves our ability to build the best possible product. It also greatly reduces the number of times we have to rework a finished feature, a frustration every engineer knows all too well.
Responses have been edited for length and clarity. Images provided by Shutterstock and listed companies.
