• Design and create of an LLM-based, multilanguage, speech-to-speech conversational AI agent for an enterprise platform that utilises client databases through RAG. Built with Model Context Protocols, LangChain, HuggingFace.
• Supervise research of a biotech startup's Machine Learning team building a digital twin platform. The portfolio includes a large protein model-based drug-target binding affinity estimator and a drug side effect prediction model.

• Led the company's Deep Learning efforts to build a state-of-the-art platform for future proof vaccine design and predictive modelling.
• Collaborated across functions to ensure mathematical and biological rigour in modelling.
• Provided strategic input in designing and utilising Deep Learning in the company's processes, transitioning them from ideas to proof-of-concept to production.
• Built the computational team as a hiring manager, growing the team from 3 to 7 people in a few months.
• Established data and software quality control standards to enable high-performing models.
• Designed and patented SARS-CoV-2 antigens, currently being tested in vivo.

• Led the IARPA-funded, multimillion-dollar FELIX project to detect signs of biological engineering in DNA sequences using bioinformatics and Deep Learning.
• Managed a team of engineers to develop an end-to-end Deep Learning pipeline for
biosecurity and was responsible for the day-to-day communication with the government.
• Developed an Automated Machine Learning (AutoML) approach for DNA sequences (DNAutoML), automating the creation of state-of-the-art Deep Convolutional Neural Networks for biology.
• Developed generative models (GANs, GRUs) to design novel synthetic E. coli promoters, resulting in several, in-silico validated new designs.
• Part of the team providing an assessment on the origins of SARS-Cov2 to the USG (featured on the IARPA website).
• Optimised the workflows to improve the detection accuracy of the biosecurity pipeline from 48% to 80% in subsequent blind evaluations.

• Co-Founder of a startup that used AI to connect scientists based on their scientific profiles.
• Led the software and algorithm development efforts and hired and managed a team of software engineers.
• Designed and built artificial intelligence algorithms to connect scientists based on their skills and experience extracted from their public scientific profile and graph matching. The algorithm provided a 22% improvement over the baseline.
• Created a generative language model to suggest scientific topics for conference organizers.
• Our company sold products and services to world-renowned institutes including EMBL EBI and the University of Cambridge.
• Exited in 2017.

• Faculty member of the Department of Computer Graphics and Image Processing researching and teaching Biomedical Image Processing and Machine Learning.
• I was the youngest Ph.D. graduate and assistant professor-appointee and one of the 10 most cited scientists in the history of the faculty.
• Initiated and managed scientific projects involving 10s of BSc, MSc, PhD students, and other junior scientists.
• Received funding from Microsoft Research, NVIDIA, and the Hungarian Academy of Sciences.

• Processed terabytes of high-throughput/high-content screening imaging data using Machine Learning and Image Processing.
• Designed and built a visual analytics tool in Java for large-scale high-content microscopy datasets (Mineotaur)
• Participated in developing the largest biomedical image resource to data, storing terabytes of data (Image Data Repository).