CTO Sessions: Dr. Rashid Mansoor, Hadean

What makes an effective tech strategy? “An effective tech strategy is one that manages to not compromise short-term profitability with long-term impact.”

IDGConnect_ctosessions_suppliedart_rashidmansoorhadean_1200x800
Hadean

Name: Dr. Rashid Mansoor

Company: Hadean

Job title: CTO & Co-Founder

Date started current role: October 2015

Location: London

Rashid is the Chief Technology Officer and Co-Founder of Hadean, a deep-tech distributed computing company based in London, UK. A member of the inaugural Entrepreneur First cohort, he previously founded VC-based data intelligence startup, Adbrain. Today, Mansoor is the driving force behind the technical innovation at Hadean.

What was your first job? While it may not quite quantify as a job, I started developing software applications for businesses when I was still in my early teens. After finishing school, I immediately started my first company developing positioning devices based on GSM technology as a cheaper alternative to GPS. I primarily fitted these to corporate vehicles for fleet management purposes. I also developed an application so you could track the devices you own and manage your fleet.

Did you always want to work in IT? My earliest interest was in paleontology. I wanted to spend my time hunting for dinosaur fossils! I really got cursorily into tech thanks to some old Apple II’s at school loaded with Apple Basic. Later I got a 386DX and spent most of my time trying to probe its inner workings and also writing code. I focused more on low level software. I started with BASIC and quickly exhausted its limits, so I moved on to Turbo Pascal and then C and assembly language. I also found Welstead’s book on neural networks around this time. I found the theoretical grounding vin that book very insightful. It helped me connect my understanding of AI to its mathematical underpinnings as well as to limitations of the hardware architectures of the time. I became particularly interested in the quest for massive parallelism around this time.

What was your education? Do you hold any certifications? What are they? I ended up doing two separate bachelor’s degrees. The first was in Computer Science (CS) but so much about the workings of hardware was relegated to “magic “that I started a parallel degree in Electrical & Electronic Engineering (EEE). I had to juggle lectures and choose which ones to attend to make it work. Naturally I went to more of the EEE ones as I had far less of a background in it and it’s not nearly as easy to self-study the topic without a well-stocked engineering laboratory.

I had a strong interest in pure mathematics from an early age and expected to see more immersion into mathematical topics in both degrees but found that this was not the case. I ended up teaching myself a lot of what you might learn in a maths degree. For my engineering degree my final year thesis was very mathematical. Apparently, the examining board questioned whether it ought to qualify as engineering. They gave me an award for it in the end… And full scholarship plus stipend for my doctoral work where I focused on further applied mathematical research.

Explain your career path. Did you take any detours? If so, discuss. After exiting my first company I decided to focus more on finishing my education. I did take some smaller detours along the way but mainly focused on learning as much as I could. I had a keen interest in robotics as well as microelectronics. I did many of my projects in robotics, but I was always interested in autonomous agents rather than industrial robotics, and this was (and still is) somewhat of a nascent field. Microelectronics had more direct relevance as I had become rather interested in the paradigm of analogue computing, particularly for applications to neural networks. It’s too complex a topic to get into here, but analogue computing offers significant benefits in power, performance and density for many applications, especially neural networks, but it is also a lot harder to do. Digital is relatively straightforward. Analogue and Mixed-Signal microelectronics is very much a black art.

I did work for a few companies along the way at various points. Some were in software, others in hardware design. Usually I did while working on research or a startup idea on the side.

I ended up taking a sabbatical in my last year of doctoral work to start my previous company, Adbrain. We built and shipped a Mobile DSP (Demand Side Platform) in about three months and subsequently built and scaled our big data platform (both real-time and batch) to tens of thousands of queries per second. The company grew to about 50 people across London, New York and the West Coast. It was while building Adbrain that I gained some of the last bits of insight for Hadean and saw the commercial opportunity there. That led to my present focus.

What type of CTO are you? I’m more of a generalist tech entrepreneur than a career CTO. I like to think I’m good at adapting, but my strength is perhaps using my broad and deep understanding of technology to predict macro-market movements. I’ve generally gone after deep tech propositions that change the status quo. My technology ideas always start with a thesis around an industry shift and is driven by that vision. I then figure out the scientific or technological principles that would make it possible to create it. From there it’s about starting a company, raising investment, hiring a team, and going to work on it.

Which emerging technology are you most excited about the prospect of? I am of course very interested in the ongoing shift to spatial computing. I do believe virtual worlds and environments, whether for lifestyle, entertainment, or other commercial reasons are set to breakout in a big way. Covid19 is a catalyst for this, but this was coming anyway. In the next 5 years I believe we will see a complete transformation in how we work, learn, socialise, and play.

The current trends in AI is fascinating, although I do think general AI or AGI is still a bit of a way away. I am interested in two areas here – Generative Adversarial Networks (GANs) which are already becoming a trend, and another long-held thesis of my mine that I’d like to work on someday: hybrid architectures recurrent dynamics and self-organising maps. There’s some more recent evidence to support this but I have some very specific ideas in mind that I’d like to get the time to work on eventually.
Are there any technologies which you think are overhyped? Why? As much as I think our current progress in AI is great, I think the hype often exceeds reality. The safety concern for one, gets blown out of proportion. Series researchers working on controlling AI are certainly moving in the right direction, but there’s often too much speculation about the imminence of apocalyptic AI. For AI to approach human intelligence, even if we figured out the neural network architectures and technology, we are still at least one hardware revolution away. We have not yet figured out either, however. Deep Learning led to a great improvement in AI’s utility for real-world problems but we are still far behind having something general enough to compete with the human brain. The hardware is not nearly there either. AGI won’t happen on silicon. The information density is just not there yet.

What is one unique initiative that you’ve employed over the last 12 months that you’re really proud of? The way in which we involve our enterprise customers in the solutions design process has been one that I’ve found to be particularly effective. As a deep tech company Hadean has some very powerful capability that in its raw form isn’t immediately applicable to your average user. It’s hard to specialise it to each unique need. Instead of approaching the table as providing an out-of-the-box solution in a throw-over-the-wall manner, approaching every customer problem as a problem that may benefit from Hadean has been effective. We’ve been able to qualify early whether Hadean is the right fit a particular customer’s needs, and within about 2 meetings we have a pretty good outline of how our technology married with that of the customer’s could lead to transformative capability. This has helped us on some of the biggest projects we’re currently working on, including gaming, virtual events, and government defence.

Are you leading a digital transformation? If so, does it emphasise customer experience and revenue growth or operational efficiency? If both, how do you balance the two? Our technology tends to be applicable to existing digital companies that are looking to significantly scale up or expand their capabilities. That said, we’re active in two areas: the first is moving conventional events to the virtual sphere while eliminating all the scale limitations of current virtual worlds, such as complexity, attendee counts and realism. It will allow us to make the metaverse a reality, with the same massive scale virtual worlds being built for online gaming and social platforms. The second is the creation of single synthetic environments which have practical uses across a range of sectors from healthcare to the military. It will enable realistic replicas of the world (often referred to as digital twins) that can be used to provide insightful planning and decision making. For example, by accurately recreating a city and its inhabitants, you could model the impact of anything from traffic management and construction projects, through to the impact of natural disasters and the spread of disease.

What is the biggest issue that you’re helping customers with at the moment? The biggest issue tends to be balancing the need for change with the dependency on legacy technology. Most of the companies we are working with are trying to scale and better embrace the limitless scaling potential of the cloud to power their products and service offerings into the next era. There is often a balance to be struck given the natural tendency to delay these shifts which tends to be at odds with the need to expand reach or capability.

How do you align your technology use to meet business goals? This is where finding a way to work out low cost ways to transition tend to be helpful. This allows immediate bottlenecks to be addressed without our customers’ existing investments being rendered obsolete. Once the evidence for what is possible is tangibly demonstrated, however, there tends to be significantly greater appetite to make a more dramatic shift.

Do you have any trouble matching product/service strategy with tech strategy? This is always a challenge as the optimal way to build technology is to take a long-term view, but this is at odds with the need to deliver. Often finding shorter-term architectures or solutions to problems while leaving the door open to replace it with a better version in the future helps steer this. Later, the work to replace a partial solution with a better alternative is motivated by a new customer need where the upside justifies the investment of time and resources.

What makes an effective tech strategy? An effective tech strategy is one that manages to not compromise short-term profitability with long-term impact. When you’re driving towards a big vision it is easy to so many compromises for the short-term that you never approach your North Star. Your North Star is what will deliver your ultimate value, whether that’s profits or fundamentally changing the way the world does a thing. There is always compromises in striking this balance, but not compromising one for the other is key. Naturally, the people aspect of this needs to be considered. The ability for the team to continue to execute under these difficult demands.

What predictions do you have for the role of the CTO in the future? That’s not an easy question to answer due to the immense diversity in the types of company, stage, technology, and market. The one constant will be the need to adapt and change to the landscape as increased automation alters traditional job responsibilities and increases opportunities for disruption. This is especially true as we stand on the cusp of the sort of AI capability that can democratise technology even further but equally may be maladaptive if we do not move quickly enough to embrace the change.

What has been your greatest career achievement? Founding Hadean based on my past experiences and subsequent insights, and later seeing the vision for Hadean realised in the form of our scale and performance capabilities. We’ve achieved something that was once considered impossible, but we proved it numerous times with our clients and also very publicly during our Eve Aether Wars tests and our protein modelling work with the Francis Crick Institute. 

Looking back with 20:20 hindsight, what would you have done differently? Too numerous to count, but if I were to pick one area, it would have been to narrow down on the simulation vertical sooner at the outset and focused on data later. It was simply quicker to make progress in simulation while there was much fatigue in big data at the time we founded.

What are you reading now? Nothing that would excite the general reader. Some recent patent work has got me thinking again about Clifford algebra so all my extracurricular reading as of late has been on Grassmann and Clifford algebras and Spinors. There’s an entire approach to physics that unifies the mathematics behind it. This is of particular value to physics systems in computer simulations (and for modelling light in graphics rendering as well) but is largely an overlooked area.

Most people don't know that I… Can be a bit eccentric.

In my spare time, I like to…Drink tea, cook or do BBQs.

Ask me to do anything but… Tedious administrative tasks.