I coach competitive volleyball throughout the year. And a saying that often comes up regardless of a win or a loss is: what got us here isn't necessarily what will get us there. To me, this is a reminder about the importance of reflecting on what you've done as you look to what's ahead. It's a special state of awareness that informs how you test, iterate, and improve upon what's working and what's not, regardless of circumstances. Those circumstances may be highs and lows in the global markets, industry trends, technology breakthroughs, team dynamics, or how well you slept the night before.
2022 was chock full of all kinds of circumstances for the inhabitants of planet Earth. 2023 -— like any new year — still has that fresh new year smell to it and is full of opportunity. This post takes a moment to look at the things that got us here and the things that will get us there.
Looking back at 2022
I’m proud of all the work the Nextmv team accomplished last year. We were a mighty team with a big list to tackle. And we wouldn’t have succeeded were it not for our customers and partners. We are endlessly thankful for them, and impressed and enriched by their creativity, engagement, and drive.
In no particular order, here are some of the highlights and notables from 2022.
The next-generation Nextmv platform
The Nextmv platform is the result of the entire company refining and building upon our existing technology stack to make it easier to access, easier to get started with, and more flexible to use across teams and use cases. In less than 5 minutes, you can go from installing Nextmv to getting the results of a pre-built decision model. From there you can create custom constraints for a food retail problem, create routes that avoid pathways like bridges and rivers, create a new value function for varying vehicle/worker types, create a model for optimizing PagerDuty schedules, and more.
MIP and Decision Diagram solving with Nextmv
Nextmv is a platform designed to model and solve many types of optimization problems using different solving paradigms, and not just different solvers that fall within a paradigm. Nextmv's origins are in state-based search built on novel decision diagram techniques. In 2021, we expanded to heuristic techniques with ALNS. In 2022, we expanded our framework to include MIP. As a result, it’s possible to build and run a customizable MIP model in minutes just as you would a decision diagram-based model with Nextmv.
Do you use an optimization solver? Do you like it?
Ryan, our CTO, posted to Hacker News to brain pick the community about solver experiences. The response was strong (it hovered in the top 10 on the homepage for a moment) and the resulting conversation was engaging. It reaffirmed to us that optimization technology and problems are often challenging and prevalent, but can fly under the radar. It also motivated us to keep driving toward our vision of making optimization technology easier to use and more accessible to more users and teams.
Partnering with Onfleet
We built upon our relationship with Onfleet, establishing a more formal partnership between our teams. As a result, it's easier than ever for customers to pair the goodness of Nextmv's customizable route optimization capabilities with Onfleet's dispatch and fleet management platform.
INFORMS 2022 and team offsite in Montreal
We are a company that formed as the pandemic got started. This means that most of our 30+ employees hadn't met one another in person or really had a chance to engage in community events. 2022 was the year that changed. We had our first company offsite in Montreal, Canada (our curling prowess is depicted below) and also hosted a technical workshop and tutorial presentation at INFORMS 2022.
Looking forward for 2023
There’s a lot to be excited about ahead of us. The aforementioned mighty team is already in the thick of building and shipping new capabilities that make optimization work smoother, more collaborative, and maybe even downright delightful.
Custom decision endpoints via hosted decision apps
Toward the end of 2022, we released a beta of Nextmv Cloud to power hosting custom decision apps (alongside our private cloud offering). I'm excited about this because it addresses two real challenges in the optimization world: 1) collaboration and 2) efficiently deploying decision models so they can have a real impact on operations. This article from VentureBeat captures it well:
Data scientists have been around since the 1950s — and they were individuals sitting in a basement working behind a terminal. But now that it’s a team sport, and the importance of that work is now being embedded into the fabric of the company, it’s essential that every person on the team is able to collaborate with everyone else: the data engineers, the data stewards, people that understand the data science, or analytics, or BI specialists, all the way up to DevOps and engineering.
“This is a big place that holds companies back because they’re not used to collaborating in this way,” Leff says. “Because when they take those insights, and they flip them over the wall, now you’re asking an engineer to rewrite a data science model created by a data scientist, how’s that work out, usually?”
You can learn more about how Nextmv is bridging the gap between optimization and production environments by joining us for a live techtalk on February 9.
Testing and benchmarking
"If you do not test it, it will not fly" are words to live by when it comes to operational models. Testing has been core to our product vision. We started with simple charts and graphs that illustrate the values of improving solutions over time, value function composition, and other visual comparisons of solutions. We're now making headway into exposing and demonstrating capabilities such as stress testing, scenario testing, historical testing, acceptance testing, and benchmarking.
We touched on some of what's to come in a recent testing and benchmarking techtalk.
Our multi-paradigm solving framework
We're continuing to make investments in our multi-paradigm solving framework. It's a powerful thing when you can simplify the engineering work needed to deploy the solving approach of your choosing in just a few minutes by using Nextmv. As you saw up above, last year we added a new MIP interface with HiGHS. This year, we're looking forward to the addition of more paradigms. We also welcome input from the community on what you would like to see us add next. Submit a comment to our forum or send us a message to let us know.
INFORMS Analytics 2023 and beyond
We're getting our events calendar in order for the year. We are excited to have a presence at INFORMS Analytics 2023 in Aurora, Colorado, so keep an eye out for more information about talks and workshops if you're going this year! And make sure to follow us on social (LinkedIn, Mastodon, Twitter) or get email updates to learn more about where else we plan to present, participate, and hang out in the conference and meetup world this year.
We’re excited about the there ahead of us. Here’s to what comes next and to accomplishing it alongside our teammates, customers, partners, investors, family members, and even the creatures great and small who call us hoomans.
