CV

Education

MSc Machine Learning (Distinction), September 2024 - September 2025

Relevant courses: Supervised Learning (84%), Probabilistic and Unsupervised Learning (76%), Statistical Natural Language Processing (81%), Bayesian Deep Learning (75%), Applied Deep Learning (85%), and Advanced Topics in Machine Learning (77%).

MSci Physics with Theoretical Physics, October 2020 - June 2024

Relevant courses: Advanced Classical Physics (83%), General Relativity (78%), Quantum Field Theory (67%), Unification (66%), Solid State Physics (70%), and Statistical Physics (72%).

Jiakang Chen, Sufia Hashim, and Carla Figueira de Morisson Faria. “Physics-informed neural networks for solving saddle-point equations in strong-field physics with tailored fields.” Accepted for publication in Physical Review Research, 2026. DOI · arXiv

Poster presentation, University College London, June - July 2026

Presented “Physics-informed neural networks for solving saddle-point equations in strong-field physics with tailored fields” at the International Conference on Attosecond and Free-Electron Laser Science.

Physics-Informed Neural Networks for Strong-Field Saddle-Point Equations, September 2025 - Present

Developed an unsupervised physics-informed neural network framework for solving complex-time saddle-point equations governing direct above-threshold ionisation in tailored laser fields.
Introduced a window parametrisation strategy that guides optimisation towards physically relevant saddle points and improves convergence stability.
Benchmarked the neural-network solutions against conventional root-finding methods across a range of laser and photoelectron-momentum parameters.
Used the predicted saddle points to calculate coherent above-threshold-ionisation photoelectron momentum distributions.

MSc Project - Mesh-free Neural Solvers for PDEs, January - April 2025

Combined machine-learning techniques with physics for solving partial differential equations.
Compared physics-informed neural networks, the deep Ritz method, and weak adversarial networks on Poisson problems up to five dimensions and on one- and two-dimensional Schrödinger equations.
Extended the framework to a laser-driven Schrödinger equation using the Kramers-Henneberger transformation.

NLP Project - Retrieval-Augmented Generation with Knowledge Graphs, January - April 2025

Proposed a retrieval-augmented generation framework integrating knowledge-graph-based retrieval and query decomposition to improve multi-hop question answering.
Constructed a domain-specific knowledge graph from unstructured text and combined dense retrieval with knowledge-graph reasoning.
Evaluated retrieval precision and answer quality and identified limitations associated with embeddings and query structuring.

Applied Deep Learning Project - Weakly Supervised Semantic Segmentation, January - April 2025

Benchmarked CAM, ECS-CAM, and CCAM class-activation-map-based localisation methods.
Designed an ensemble framework with conditional random field refinement to generate higher-quality pseudo-labels.
Trained a U-Net using the pseudo-labels, achieving 76% IoU compared with a 68% fully supervised baseline.

MSci Project - Black Holes and Branes in Supergravity, July 2023 - April 2024

Derived brane solutions in (D=11) and other dimensions using advanced techniques in general relativity and supergravity.
Developed an understanding of spinor fields in higher-dimensional spacetimes.
Investigated T-duality in string theory and related dimensional-reduction methods; achieved a project mark of 78.05%.

Programming: Python
Machine Learning: PyTorch, scikit-learn, physics-informed neural networks, deep learning, probabilistic modelling
Scientific Computing: NumPy, SciPy, Pandas, Matplotlib, numerical optimisation, root finding, differential equations
Tools: Git, GitHub, Jupyter, Visual Studio Code, Mathematica, LaTeX