Acquired high-dimensional neural activity data and used machine learning to model the information about social status encoded in the rodent hypothalamus.o Construction of recording setups for synchronised acquisition of neural and behavioural data, including real-time computer vision workflows using a reactive programming approach.o Developed data cleaning and pre-processing pipelines across diverse types of data, from continuous timeseries to lists of annotations with timestamps and converting them into compatible data structures.o Developed custom toolkits for common operations in multi-modal datasets, such as data segmentation based on features of one of the modalities.o Developed data validation pipelines, e.g. to ensure correct synchronisation of cross-modal datasets.o Used GLMs to dynamically predict animal behaviour e.g. whether a mouse will win a competitive encounter against another mouse from brain activity data.o Used unsupervised and supervised dimensionality reduction methods including autoencoders for denoising, feature selection, and visualisation of complex datasets.o Used support vector machines for the classification of the mouse’s own as well as the opponents social rank from brain activity.o Hands-on experience with reinforcement learning to model the emergence of hierarchical social behaviour among multiple agents in a resource constrained environment.o Design and management of recording databases and automated filtering of datasets based on those records.o Extensive use of inferential statistics for hypothesis testing on highly varied data types.o Stereotactic surgeryo Histology technique (RNAScope, immunohistochemistry)o Confocal microscopyo Conducting behavioural assays and simultaneous neural recordingso Collaborated and provided analytical toolkits for other projects in the group.o Trained colleagues in machine learning methods.o Mentored a master’s student.

Honours project analysing whole-brain calcium imaging data in Zebrafish that compared the functional differences between fish with distinct motor behaviour profiles that indicated trait differences in anxiety.- Behavioural profiling- 3D registration of all brains to a reference dataset- Identification of brain areas with differential neural activation based on behavioural profile- Identification of differences in whole-brain functional connectivity based on behavioural profile using Bayesian functional connectivity inference.

Identified novel cellular structures in electron microscopic images of the developing cerebellum - The tunnelling nanotubes previously only observed in cultured neurons, but not in vivo. Worked with 3DS Max to create visualisations of the volumetrically reconstructed cells.Won the best poster award at the concluding internship symposium.

Built Matlab tools for the analysis of the structural features of 3D mitral neuron reconstructions from electron microscope images of the olfactory bulb.

Laboratory assistant in a blood diagnostics laboratory. I conducted routine blood tests and maintained appropriate equipment and test result records for reporting to clinicians.