Biography

Building the human-evidence layer for drug discovery.

Founding engineer at a pre-seed applying AI to drug manufacturing. Primarily centered around document analysis and FDA compliance.

Support integration and analysis of multi-modal clinical data. Architected and implemented a unified framework to simplify curating tabular and tensor data from national biobanks, reducing data scientists' wrangling time from weeks to hours. Centralised data repositories and built preprocessing pipelines for radiology data to support medical imaging research at biobank scale. Embedded with ML scientists to collaboratively design and validate insitro's ML-toolkit through real analyses. Developed ML models for cell slide quality control. Promoted to Lead Engineer.

Responsible for cloud compute infrastructure. Set up a cloud notebook platform (JupyterHub), and rolled out self-serve notebooks to ~200 users with tracking, auto-scaling, and cost-optimization. Organised centralized repositories for internal and normalized external RNA-Seq data for scientists, engineers, and researchers. Built full-stack data explorers for scientists, computational biologists, and statistical geneticists. Partnered with neuroscientists to analyze perturbation sequencing screens and optimize wet-lab protocols. This role bridged cloud-compute infrastructure with target-discovery efforts.

Led 'Datahub' - an initiative to centralise and organise data infrastructure for the Biohub. Designed and built Metahub, an in-house LIMS platform to support diverse research across the Biohub, modeling heterogeneous metadata with a graph database (Neo4J). Developed a full-stack website for lab data input and tracking with the genomics platform. Set up and orchestrated Nextflow pipelines for genomics, transcriptomics, and proteomics - that were also later contributed to OpenPipelines. Set up data portals for cell atlases. Completed ad-hoc single-cell analyses and mentored two interns.

Bioinformatics engineer within the Population Genomics group. Helped develop a tool for merging and storing terabyte-scale variant call files (VCFs) into a single gVCF store. Created software to help triage bugs and locate algorithmic bottlenecks in the platform. Provided exposure to industry-scale genomics infrastructure.

Bioinformatician at a genomics company. Developed an exome analysis pipeline for variant calling as part of the 1 Million Genomes Project. Built a website that enabled geneticists to review variants, investigate causative mutations, and auto-generate reports for participants. It consulted public data sources and helped visualize the exome analysis results. First experience building Nextflow pipelines and combining bioinformatics with full-stack development.

Software developer at a Berlin-based CPaaS company. Developed a pricing API for communication services and a web-scraping tool to collect competitor prices, both consumed by internal services. Deployed services with Docker and AWS. Part-time role during university that transitioned to full-time, providing early experience with backend development, SQL databases, and cloud infrastructure.

R&D internship at an AI-driven sales coaching startup. Researched and developed a hardware device to support real-time call assistance. Sourced electrical components, soldered and tested boards, and designed control software to manipulate audio on microcontrollers. Built a web app to communicate between the ML backend and audio data. Gap year role combining hardware prototyping with software development.

Engineering internship at an early-stage startup (5 people) developing real-time pathogen testing for food safety using a phage-based magneto-elastic biosensor. Contributed across data analysis (signal smoothing, visualization GUIs), hardware programming (coil-winder control software, prototype lighting), and business operations. First internship during gap year, introducing me to multidisciplinary teamwork and coding.