The People Behind SMART M3S: Dr Alok Prakash, Senior Scientific Director

Dr Alok Prakash shared perspectives on AI readiness and the future of smart cities in an increasingly connected world at the ATxEnterprise 2026 event.

Growing up in Bokaro Steel City in India, Dr Alok Prakash was fascinated by repairing broken gadgets.  In a town where most households repaired what broke and stretched what they had, he would hunt for broken radios and cassette players, take them apart and try to bring them back to life, which sparked an early fascination with doing more with less. That curiosity carried him to Electronics Engineering at the Indian Institute of Technology (Indian School of Mines) Dhanbad (IIT (ISM) Dhanbad), and later to Nanyang Technological University (NTU Singapore), where his PhD focused on getting small, low-power computing platforms to do ambitious things.

Today, as Senior Scientific Director of SMART’s Mens, Manus, and Machina (M3S) interdisciplinary research group (IRG), Dr Alok works across embodied AI, LLM agents, AI for learning and the future of work. He leads the programme’s scientific direction while mentoring postdocs and research engineers and working closely with the IRG’s Principal Investigators. Across every project, whether it's optimising AI agents or studying workplace automation, he remains guided by the same question he first asked as a child: how can we accomplish more with less?

In this feature, Dr Alok shares how his love for science fiction and tinkering with broken gadgets in his youth shaped his research approach, what keeps him motivated across diverse AI challenges, and why his early childhood experiences remain the philosophy guiding his work today.

What do you do at SMART M3S?

I work with our research teams across embodied AI, LLM agents, AI for learning and the future of work to shape the overall scientific direction, recruit and mentor our postdocs and research engineers, and coordinate closely with the IRG’s Principal Investigators. 

As Senior Scientific Director, I also set the direction for how the programme equips itself for AI and Robotics research: how we access the computing power we need, and how we safeguard our partners' data, which often determines what research is feasible at all. 

Keeping a research programme at the frontier while finding ways to maximise the impact of every resource is its own kind of engineering challenge. A large part of my time also goes into working with Singapore’s government agencies, industry partners and community organisations, so that our research benefits Singapore and the world. That is something I care about personally. One area I am personally closest to at the moment is our future-of-work research, where we study how AI and automation reshape jobs at the level of individual tasks, working with colleagues at MIT and national partners such as the Skills and Workforce Development Agency (SWDA).

How did you first become interested in your field of work, and what motivated you to pursue it as a career then?

Where I grew up, most households repaired what broke and stretched what they had, so doing more with less was simply everyday life. I just happened to enjoy tinkering with gadgets more than most. That fascination led me to Electronics Engineering at IIT (ISM) Dhanbad, and later to work at IBM Bangalore. 

My path to Singapore was actually quite unexpected. I had an admission offer from the University of Michigan, but pursuing it would have meant taking a loan far beyond what my family could comfortably manage. A friend interning at NTU Singapore suggested I explore Singapore instead, and what began as an unexpected detour became the best accident of my life. I moved here in 2007 to pursue a PhD at NTU Singapore, where I worked on getting small, low-power computing platforms to do ambitious things. Nearly half my life later, that same idea continues to shape my research: how do you accomplish more with less?

What is the biggest motivation that keeps you going at work now, especially during the hard/challenging times?

Honestly, it is the work itself. I am happiest when I am close to the actual research rather than floating above it. I deeply appreciate my current role at M3S, which gives me the opportunity to dive into several very different problems in the same week: how an AI agent should manage its memory, how a robot learns a task it could not do a month earlier, whether an AI tutor genuinely helps a student, or how a job breaks down into tasks that AI can and cannot take over. Some of this is fundamental research that may take years to translate into applications, while other projects are already being tested with partners; I sincerely enjoy both aspects. 

While my days and weeks often run long, the questions I enjoy most all seem to need the same combination: a willingness to dive all the way down into the technical details, and the chance to zoom out to envision the big picture. I feel most useful at that junction, and on challenging weeks, that is what keeps me going. 

Another aspect is that the diversity of research in M3S helps to keep things fresh. If I get stuck on one project, there is always something else to work on, giving me time to return to the original problem with a different perspective.

How does your work benefit the society/research communities?

We hold ourselves to a simple standard: prove that something works before scaling it up because I believe AI should not be deployed without proper evidence. For example, we recently completed a randomised controlled trial of AI-assisted tutoring with a university in Peru and are now analysing the results, with a similar planned study with MENDAKI in Singapore. The same thinking shaped my earlier work on transport sensors that we piloted round the clock on Singapore's roads, and my contribution to the national Traffic Probe Data Standard.

More recently, we have been mapping Singapore’s national skills frameworks onto a detailed ontology of work activities, so that conversations about reskilling can be grounded in evidence drawn from actual tasks rather than speculations about the impact of AI, which often fuels more anxiety than insight. If our work helps Singapore adopt AI carefully as well as quickly, I would consider that a real contribution.

Dr Alok Prakash speaking at a workshop hosted by M3S attended by senior decision-makers across Singapore’s education and workforce sectors.

Could you share any highlights or important projects/milestones that you have worked on/achieved at SMART M3S?

The past two years have been among the most productive of my career. Our efficient-AI research line with MIT produced MEM1 (ICLR 2026), which enables AI agents to work on long tasks without requiring ever-increasing amounts of memory - this research received the ‘Best Paper Award’ from over 230 submissions at the MTI-LLM workshop in NeurIPS 2025. 

We also published TETRIS (ACL 2025) and MineDraft (ICML 2026), which make large language models faster and cheaper to run, and a deep ontology of work activities that now underpins our skills research with Singapore’s agencies. Beyond the publications, our National Research Foundation (NRF) programme on biologically inspired embodied AI with Agency for Science, Technology and Research (A*STAR) I2R, where I serve as a Team-PI, also kicked off this year. 

Most recently in July, I also convened a closed-door national workshop where senior leaders from government, universities, industry and community organisations spent a day discussing a lifetime of learning in the AI era.

What do you hope to see happen in your research field in the next 10 years and what excites you most about the possibilities ahead?

I hope that over the next decade, AI becomes something that adapts to people rather than the other way around, be it helping a child learn better, a worker moving between careers, or a senior staying independent for longer. I also expect embodied AI to finally leave the lab, and that is exactly why efficiency matters so much to me, because intelligence only becomes useful when it fits within the power, cost and safety budgets of the real world. What excites me most is that Singapore, where researchers, agencies and industry genuinely trust one another, is probably the best place in the world to test this future first.

Balancing work and personal life can be challenging. What are your current hobbies to relax and recharge?

I am a huge fan of pens, especially fountain pens. My first thinking still happens on paper; I like to handwrite my notes and still do so every single day, including meetings. I can happily spend a weekend afternoon just cleaning and refilling my pens. 

My other indulgence is gadgets. I read reviews and watch videos for practically every new phone, laptop or wearable that launches. Back in my younger days, I used to change my phone almost every month with help from Carousell; these days I have learnt some restraint and settled for just one change a year!

Dr Alok Prakash with a Japan Railways (JR) Hokkaido conductor during his Japan trip in 2026. 

Can you share one thing that you’d love to do on your bucket list?

Taking train journeys around the world. I have been a fan of trains for as long as I can remember, and trips across Japan in 2019 and 2026 and parts of Canada in 2023 only deepened it. The dream is to keep adding railways to the list: the Swiss passes, the Trans-Siberian, the mountain lines back in India. For someone who has spent years doing transport research, I suppose it is fitting that my idea of a perfect holiday is simply being a passenger!