Meet the ADEPT Decisions Management Team: Michael Lorenat

Introduction

The entire ADEPT Decisions team has always aimed to provide the ‘personal touch’ to our clients; whether it means going the extra mile on a project, accommodating short-notice configuration changes, or just providing a sounding board to the credit risk team.

We believe that this flexible, ‘can do’ attitude is one of the aspects of ADEPT Decisions which set us apart from larger and less accommodating corporate vendors.

With this in mind, we decided it was a suitable time to shine a light on the unique personalities behind the ADEPT Decisions software. This is a management team which prides itself on delivering above and beyond the call of duty!

Michael Lorenat

Describe your role at ADEPT Decisions.
I am the CTO of ADEPT Decisions, and the software architect of the core Decision Engine of the ADEPT Decisions Platform – the “D” in “ADP”. Besides evolving improvements to the engine, I am also responsible for consulting on improvements to the domain and subscriber-specific configurations of the decision processes and components running in the engine.

How long have you been with the company?
I started with the company at its inception in 2015. I was also the original architect and development manager for the Python and Flash implemented first generation ADEPT decision engine, which was launched in 2004.

What led you to join ADEPT Decisions?
The first-generation ADEPT was partly developed and initially marketed in Thailand. There were successes, and there were significant challenges. After securing rights to the first-generation source code, Stephen John Leonard founded ADEPT Decisions and asked me to the join the effort. This was a wonderful opportunity to engage with an entirely new market: Africa. More importantly, it was an opportunity to help create ADP, which introduced configurable data processing capabilities and offered a fully cloud-hosted product.

What are the biggest challenges you have faced so far?
With respect to decision engines, over the years, machine learning has come to dominate predictive modelling. With a data scientist mindset, it just takes a little bit of code to, for example, consume a probability prediction, a risk score or a cluster identifier and transform it to accept/decline/review decisions along with appropriate value settings of the attributes of a particular credit offer. In contrast, a business user configurable decision engine keeps a business user and their perspectives in the midst of the decision-making. Provided that the evolving business decisions are easy and fast to configure, deploy, and evaluate.

It has been a challenge to evolve technologies while keeping the product offerings relevant and attractive to the ever-changing market.

What is your background?
I was a psychology BA who visited an employment agency to help find a job during a deep recession. The agent looked at me sceptically until he discovered that I was a competent typist and adequate 10-key operator. So, I was able to land a job as a minimum wage clerk in a brokerage house in San Francisco – a universe away from the excitement and promise of Silicon Valley. Eventually, I was able to parlay a single college Fortran programming class into a programmer job at only a little better than minimum wage. I caught the bug and went on to get a Master’s and eventually a PhD in computer science focused on Human Factors. I landed a job with Fair, Isaac and worked there for eight years helping to bring visual decision design eventually into the FICO Blaze software. In 2004, I co-founded Adaptivate to help develop a new visual decision engine for smaller financial markets ignored by FICO at the time.

How did you start in the industry?
I learned about credit risk, modelling and scorecards while trying to serve the analysts at FICO before they were labelled data scientists. I started prototyping visual decision engines with Visual Basic and later led a development team in creating modelling software.

What changes have you seen over the years?
My first piece of software was written in Basic typed into a machine which created a paper tape. Nearly ten years later, I was using punch cards. I was telecommuting with a teletype machine a year or two later. A few years later, I started using Windows 2, which offered tiled – not overlapping windows. It is mind-boggling how quickly technology evolved. Now, I have an iPhone with vastly more computing power than the climate-controlled room full of CDC mainframes I once programmed. In the cloud, I’m occasionally provisioning Kubernetes pods on Azure and AWS as routinely as I once used a black Sharpie to mark across the edges of a deck of punch cards to help recover from a catastrophic deck fall.

What advice would you give to those starting out in the industry?
Ideally, today, someone starting out should strive to gain experience and understanding in three areas: the business side of credit risk management, data analysis and modelling, and software development. As you grow to understand your aptitudes and your interests, you can focus on one of the areas; but maintaining a broad perspective will help you collaborate with your colleagues more effectively.

The other perhaps more important piece of advice is to establish and cultivate relationships. In the early stages of your career, mentors are crucial.

What do you like to do in your spare time?
My wife and I recently moved from the wine country of Napa Valley (not Silicon Valley) to the wine country of Southern Oregon. Here I enjoy hiking and biking and a bit of golf.

What’s your favourite quote?
I annoyed my young daughters with a frequent after school question: “Did you ask any good questions today?” Via Google, I have found that Physicist Isador Isaac Rabi attributed this quote to his mother. So, it is possible I did not invent this independently.

What’s a little-known fact about you?
I attended the 61st Academy Awards in 1989. Unfortunately, my wife did not win an Oscar.

About the Author
Michael Lorenat is CTO of ADEPT Decisions and has held a wide range of roles in software and the credit risk industry since 1980.