Dr. Mike Flaxman is at present the VP of Product at HEAVY.AI, having beforehand served as Product Supervisor and led the Spatial Information Science apply in Skilled Providers. He has spent the final 20 years working in spatial environmental planning. Previous to HEAVY.AI, he based Geodesign Technolgoies, Inc and cofounded GeoAdaptive LLC, two startups making use of spatial evaluation applied sciences to planning. Earlier than startup life, he was a professor of planning at MIT and Business Supervisor at ESRI.
HEAVY.AI is a hardware-accelerated platform for real-time, high-impact knowledge analytics. It leverages each GPU and CPU processing to question huge datasets shortly, with help for SQL and geospatial knowledge. The platform consists of visible analytics instruments for interactive dashboards, cross-filtering, and scalable knowledge visualizations, enabling environment friendly huge knowledge evaluation throughout varied industries.
Are you able to inform us about your skilled background and what led you to hitch HEAVY.AI?
Earlier than becoming a member of HEAVY.AI, I spent years in academia, in the end educating spatial analytics at MIT. I additionally ran a small consulting agency, with a wide range of public sector purchasers. I’ve been concerned in GIS tasks throughout 17 international locations. My work has taken me from advising organizations just like the Inter American Growth Financial institution to managing GIS know-how for structure, engineering and building at ESRI, the world’s largest GIS developer
I keep in mind vividly my first encounter with what’s now HEAVY.AI, which was when as a guide I used to be chargeable for state of affairs planning for the Florida Seashores Habitat Conservation Program. My colleagues and I had been struggling to mannequin sea turtle habitat utilizing 30m Landsat knowledge and a pal pointed me to some model new and really related knowledge – 5cm LiDAR. It was precisely what we would have liked scientifically, however one thing like 3600 occasions bigger than what we’d deliberate to make use of. For sure, nobody was going to extend my funds by even a fraction of that quantity. In order that day I put down the instruments I’d been utilizing and educating for a number of a long time and went in search of one thing new. HEAVY.AI sliced by and rendered that knowledge so easily and effortlessly that I used to be immediately hooked.
Quick ahead a number of years, and I nonetheless assume what HEAVY.AI does is fairly distinctive and its early guess on GPU-analytics was precisely the place the trade nonetheless must go. HEAVY.AI is firmly focussed on democratizing entry to huge knowledge. This has the info quantity and processing pace part after all, basically giving everybody their very own supercomputer. However an more and more essential side with the appearance of huge language fashions is in making spatial modeling accessible to many extra folks. Nowadays, moderately than spending years studying a posh interface with 1000’s of instruments, you may simply begin a dialog with HEAVY.AI within the human language of your alternative. This system not solely generates the instructions required, but additionally presents related visualizations.
Behind the scenes, delivering ease of use is after all very tough. Presently, because the VP of Product Administration at HEAVY.AI, I am closely concerned in figuring out which options and capabilities we prioritize for our merchandise. My in depth background in GIS permits me to actually perceive the wants of our clients and information our improvement roadmap accordingly.
How has your earlier expertise in spatial environmental planning and startups influenced your work at HEAVY.AI?
Environmental planning is a very difficult area in that you should account for each totally different units of human wants and the pure world. The final answer I discovered early was to pair a technique referred to as participatory planning, with the applied sciences of distant sensing and GIS. Earlier than selecting a plan of motion, we’d make a number of eventualities and simulate their constructive and unfavorable impacts within the laptop utilizing visualizations. Utilizing participatory processes allow us to mix varied types of experience and remedy very complicated issues.
Whereas we don’t usually do environmental planning at HEAVY.AI, this sample nonetheless works very nicely in enterprise settings. So we assist clients assemble digital twins of key components of their enterprise, and we allow them to create and consider enterprise eventualities shortly.
I suppose my educating expertise has given me deep empathy for software program customers, notably of complicated software program programs. The place one pupil stumbles in a single spot is random, however the place dozens or tons of of individuals make related errors, you already know you’ve obtained a design situation. Maybe my favourite a part of software program design is taking these learnings and making use of them in designing new generations of programs.
Are you able to clarify how HeavyIQ leverages pure language processing to facilitate knowledge exploration and visualization?
Nowadays it appears everybody and their brother is touting a brand new genAI mannequin, most of them forgettable clones of one another. We’ve taken a really totally different path. We consider that accuracy, reproducibility and privateness are important traits for any enterprise analytics instruments, together with these generated with massive language fashions (LLMs). So now we have constructed these into our providing at a elementary stage. For instance, we constrain mannequin inputs strictly to enterprise databases and to supply paperwork inside an enterprise safety perimeter. We additionally constrain outputs to the most recent HeavySQL and Charts. That signifies that no matter query you ask, we are going to attempt to reply along with your knowledge, and we are going to present you precisely how we derived that reply.
With these ensures in place, it issues much less to our clients precisely how we course of the queries. However behind the scenes, one other essential distinction relative to client genAI is that we effective tune fashions extensively towards the particular sorts of questions enterprise customers ask of enterprise knowledge, together with spatial knowledge. So for instance our mannequin is great at performing spatial and time sequence joins, which aren’t in classical SQL benchmarks however our customers use every day.
We package deal these core capabilities right into a Pocket book interface we name HeavyIQ. IQ is about making knowledge exploration and visualization as intuitive as attainable through the use of pure language processing (NLP). You ask a query in English—like, “What had been the climate patterns in California final week?”—and HeavyIQ interprets that into SQL queries that our GPU-accelerated database processes shortly. The outcomes are offered not simply as knowledge however as visualizations—maps, charts, no matter’s most related. It’s about enabling quick, interactive querying, particularly when coping with massive or fast-moving datasets. What’s key right here is that it’s typically not the primary query you ask, however maybe the third, that basically will get to the core perception, and HeavyIQ is designed to facilitate that deeper exploration.
What are the first advantages of utilizing HeavyIQ over conventional BI instruments for telcos, utilities, and authorities businesses?
HeavyIQ excels in environments the place you are coping with large-scale, high-velocity knowledge—precisely the sort of knowledge telcos, utilities, and authorities businesses deal with. Conventional enterprise intelligence instruments typically battle with the quantity and pace of this knowledge. As an illustration, in telecommunications, you might need billions of name data, however it’s the tiny fraction of dropped calls that you should concentrate on. HeavyIQ lets you sift by that knowledge 10 to 100 occasions quicker due to our GPU infrastructure. This pace, mixed with the power to interactively question and visualize knowledge, makes it invaluable for danger analytics in utilities or real-time state of affairs planning for presidency businesses.
The opposite benefit already alluded to above, is that spatial and temporal SQL queries are extraordinarily highly effective analytically – however might be sluggish or tough to put in writing by hand. When a system operates at what we name “the pace of curiosity” customers can ask each extra questions and extra nuanced questions. So for instance a telco engineer would possibly discover a temporal spike in gear failures from a monitoring system, have the instinct that one thing goes mistaken at a selected facility, and test this with a spatial question returning a map.
What measures are in place to forestall metadata leakage when utilizing HeavyIQ?
As described above, we’ve constructed HeavyIQ with privateness and safety at its core. This consists of not solely knowledge but additionally a number of sorts of metadata. We use column and table-level metadata extensively in figuring out which tables and columns include the data wanted to reply a question. We additionally use inside firm paperwork the place supplied to help in what is called retrieval-augmented technology (RAG). Lastly, the language fashions themselves generate additional metadata. All of those, however particularly the latter two might be of excessive enterprise sensitivity.
In contrast to third-party fashions the place your knowledge is often despatched off to exterior servers, HeavyIQ runs domestically on the identical GPU infrastructure as the remainder of our platform. This ensures that your knowledge and metadata stay underneath your management, with no danger of leakage. For organizations that require the very best ranges of safety, HeavyIQ may even be deployed in a very air-gapped setting, guaranteeing that delicate info by no means leaves particular gear.
How does HEAVY.AI obtain excessive efficiency and scalability with huge datasets utilizing GPU infrastructure?
The key sauce is actually in avoiding the info motion prevalent in different programs. At its core, this begins with a purpose-built database that is designed from the bottom as much as run on NVIDIA GPUs. We have been engaged on this for over 10 years now, and we actually consider now we have the best-in-class answer with regards to GPU-accelerated analytics.
Even the perfect CPU-based programs run out of steam nicely earlier than a middling GPU. The technique as soon as this occurs on CPU requires distributing knowledge throughout a number of cores after which a number of programs (so-called ‘horizontal scaling’). This works nicely in some contexts the place issues are much less time-critical, however usually begins getting bottlenecked on community efficiency.
Along with avoiding all of this knowledge motion on queries, we additionally keep away from it on many different widespread duties. The primary is that we will render graphics with out shifting the info. Then in order for you ML inference modeling, we once more do this with out knowledge motion. And if you happen to interrogate the info with a big language mannequin, we but once more do that with out knowledge motion. Even in case you are a knowledge scientist and need to interrogate the info from Python, we once more present strategies to do that on GPU with out knowledge motion.
What meaning in apply is that we will carry out not solely queries but additionally rendering 10 to 100 occasions quicker than conventional CPU-based databases and map servers. Once you’re coping with the huge, high-velocity datasets that our clients work with – issues like climate fashions, telecom name data, or satellite tv for pc imagery – that sort of efficiency increase is totally important.
How does HEAVY.AI keep its aggressive edge within the fast-evolving panorama of massive knowledge analytics and AI?
That is an important query, and it is one thing we take into consideration consistently. The panorama of massive knowledge analytics and AI is evolving at an extremely speedy tempo, with new breakthroughs and improvements taking place on a regular basis. It actually doesn’t harm that now we have a ten yr headstart on GPU database know-how. .
I feel the important thing for us is to remain laser-focused on our core mission – democratizing entry to huge, geospatial knowledge. Which means frequently pushing the boundaries of what is attainable with GPU-accelerated analytics, and guaranteeing our merchandise ship unparalleled efficiency and capabilities on this area. An enormous a part of that’s our ongoing funding in growing customized, fine-tuned language fashions that really perceive the nuances of spatial SQL and geospatial evaluation.
We have constructed up an in depth library of coaching knowledge, going nicely past generic benchmarks, to make sure our conversational analytics instruments can interact with customers in a pure, intuitive method. However we additionally know that know-how alone is not sufficient. We’ve got to remain deeply related to our clients and their evolving wants. On the finish of the day, our aggressive edge comes all the way down to our relentless concentrate on delivering transformative worth to our customers. We’re not simply retaining tempo with the market – we’re pushing the boundaries of what is attainable with huge knowledge and AI. And we’ll proceed to take action, irrespective of how shortly the panorama evolves.
How does HEAVY.AI help emergency response efforts by HeavyEco?
We constructed HeavyEco after we noticed a few of our largest utility clients having vital challenges merely ingesting immediately’s climate mannequin outputs, in addition to visualizing them for joint comparisons. It was taking one buyer as much as 4 hours simply to load knowledge, and if you end up up towards fast-moving excessive climate situations like fires…that’s simply not adequate.
HeavyEco is designed to supply real-time insights in high-consequence conditions, like throughout a wildfire or flood. In such eventualities, you should make selections shortly and based mostly on the absolute best knowledge. So HeavyEco serves firstly as a professionally-managed knowledge pipeline for authoritative fashions reminiscent of these from NOAA and USGS. On prime of these, HeavyEco lets you run eventualities, mannequin building-level impacts, and visualize knowledge in actual time. This offers first responders the vital info they want when it issues most. It’s about turning complicated, large-scale datasets into actionable intelligence that may information rapid decision-making.
Finally, our objective is to provide our customers the power to discover their knowledge on the pace of thought. Whether or not they’re working complicated spatial fashions, evaluating climate forecasts, or making an attempt to determine patterns in geospatial time sequence, we would like them to have the ability to do it seamlessly, with none technical obstacles getting of their method.
What distinguishes HEAVY.AI’s proprietary LLM from different third-party LLMs by way of accuracy and efficiency?
Our proprietary LLM is particularly tuned for the sorts of analytics we concentrate on—like text-to-SQL and text-to-visualization. We initially tried conventional third-party fashions, however discovered they didn’t meet the excessive accuracy necessities of our customers, who are sometimes making vital selections. So, we fine-tuned a spread of open-source fashions and examined them towards trade benchmarks.
Our LLM is far more correct for the superior SQL ideas our customers want, notably in geospatial and temporal knowledge. Moreover, as a result of it runs on our GPU infrastructure, it’s additionally safer.
Along with the built-in mannequin capabilities, we additionally present a full interactive person interface for directors and customers so as to add area or business-relevant metadata. For instance, if the bottom mannequin doesn’t carry out as anticipated, you may import or tweak column-level metadata, or add steerage info and instantly get suggestions.
How does HEAVY.AI envision the position of geospatial and temporal knowledge analytics in shaping the way forward for varied industries?
We consider geospatial and temporal knowledge analytics are going to be vital for the way forward for many industries. What we’re actually targeted on helps our clients make higher selections, quicker. Whether or not you are in telecom, utilities, or authorities, or different – being able to investigate and visualize knowledge in real-time is usually a game-changer.
Our mission is to make this type of highly effective analytics accessible to everybody, not simply the large gamers with huge sources. We need to be certain that our clients can make the most of the info they’ve, to remain forward and remedy issues as they come up. As knowledge continues to develop and change into extra complicated, we see our position as ensuring our instruments evolve proper alongside it, so our clients are at all times ready for what’s subsequent.
Thanks for the good interview, readers who want to be taught extra ought to go to HEAVY.AI.
