2025 cohort
These researchers mentored the first BSRI cohort, between early 2025 and the close of that cycle. Eight of them are mentoring again in 2026.
Computing, AI & interaction
PhD student · Purdue University
Also mentoring in 2026Applies machine learning and natural language processing across finance and civil engineering, including large language models for time-series forecasting and deepfake detection, alongside work on the seismic response of soil and embankments.
Studies how much of a person’s thinking can be read from their gaze, and uses eye tracking to build real-time assistants for immersive environments — along with the design problems such agents raise.
Investigates how people adapt technology to work around infrastructural limits, particularly in the global south, drawing on critical HCI, postcolonial computing and participatory design to make technology design more equitable.
Builds unsupervised and self-supervised methods for multi-scale generative modelling of subcellular objects, and adapts them to 2D and 3D biological images.
Applies pretrained language models — transformers and state space models — to genomics data, particularly post-transcriptional regulatory mechanisms, with earlier research and industry work in Bangla NLP.
Explores fault attacks on deep neural network accelerators across FPGA, TPU and GPU hardware.
Computation in biology & health
PhD student · University of British Columbia
Also mentoring in 2026Works on machine learning for regulatory genomics — predicting how DNA sequence determines gene expression, and what that reveals about the non-coding genome.
PhD student · University of Maryland, Baltimore County
Also mentoring in 2026Develops machine learning for biomedical signals on wearable devices, including real-time stress detection from ECG and PPG data as a scalable alternative to clinic-based mental health assessment.
Physics & astronomy
Postdoctoral fellow · Waterloo Centre for Astrophysics
Also mentoring in 2026Studies how galaxies form and evolve across cosmic time in different environments, using cosmological hydrodynamic simulations and photometric surveys to analyse large samples of galaxies and clusters — and to ask when, how and why galaxies stop forming stars.
Postdoctoral researcher · Iowa State University
Also mentoring in 2026Studies star and planet formation: how turbulence and initial mass function variations shape stellar birth, and how non-ideal magnetohydrodynamic effects govern protoplanetary disc evolution and the growth of planetary embryos.
PhD student · The Ohio State University
Also mentoring in 2026Studies how nuclear materials behave under irradiation, combining theoretical and computational modelling with hands-on experiment so that simulations can be validated against real observations.
Materials & nanotechnology
Postdoctoral researcher · Stanford University
Also mentoring in 2026Probes the physics of low-dimensional materials to build multifunctional opto-electro-mechanical nanodevices, with the goal of unconventional memory, logic and energy-harvesting devices that cut the energy cost of computation.
Models bio and soft materials and nano-microfluidic systems, combining physical chemistry, statistical physics and machine learning. Projects include polyelectrolyte-grafted nanochannels, peptide-amphiphile nanofibre formation, and siRNA payload distribution in lipid nanoparticles.
PhD candidate · University of South Carolina
Develops graph neural networks, transformers and diffusion models for materials informatics — crystal structure prediction, property prediction, and generative modelling of materials. The crystal structure problem is a close cousin of protein structure prediction.
Mechanical, energy & transport
University of Calgary
Also mentoring in 2026Uses low-value biomass — specifically aspen fibre — in fused filament fabrication to make wood-PLA biocomposites, working on the chemical treatments that let hydrophilic wood fibre bond with a hydrophobic polymer matrix, and on how much fibre a print can carry before it clogs.
Works on defrosting and deicing strategies that cut energy use and downtime, studying how frost nucleates and grows across temperature, humidity and surface wettability, and developing electro-thermal defrosting using conductive thin-film coatings.
Works on transportation safety, currently wrong-way driving detection technologies and school bus safety, using statistical analysis, machine learning and video analytics alongside survey data collection.
