Enveda biosciences is using automation and machine learning to map the massive unknown chemical space in nature with the goal of finding new therapeutics. Data science at Enveda includes:- Deep learning (transformers, LLMs, graph NNs) for molecule structure elucidation from mass spectra.- Mathematical modeling of complex chemical mixtures to identify bioactive molecular scaffolds.- Reasoning on molecular networks and massive biomedical knowledge graphs for lead generation, target ID, and prediction of downstream ADME+tox.

Taught CS501R: Deep Learning Applications in Health Care and Life Sciences

Machine learning and data science consultant using computer vision, natural language processing, graph neural networks, advanced analytics and optimization. Domains include biotech, health care, document information extraction and conservation.

Used computer vision, chemical modeling and probabilistic analysis to identify the biological effects of rare diseases on fluorescence images of human cells, and screened for potential treatments that reverse those effects. Combined this with the use of graph neural networks to probe the relationships between the chemical structure of thousands of potential drugs and downstream biological effects.Contributed to the discovery and investigation of three drug candidates currently in clinical trials for treatment of rare genetic diseases.

Wrote Python application for creating, storing, and deploying machine learning models for healthcare providers. Contributor to open source healthcare.ai

As the lead data scientist on the company's flagship predictive product, NeuralView, I conducted research and built machine learning models and applications in Python and Java (neural networks, random forests, clustering, LDA, etc), as well as developed predictive analytics models around customer lifecycle.I also gained experience interfacing with enterprise clients to determine needs, conduct A/B tests and POCs, coordinate research efforts, educate and present results.

Single-cell sociobiology, evolutionary game theory, population dynamics, stochastic gene expression, social decision-making.

Analyzed geographically-constrained human social network structures and their relationships to social capital, social cohesion, and population resilience to natural disaster.Using Twitter data, modeled social unrest and rioting in Latin America as loss of resilience in a dynamical system with early warning signals related to critical slowing down (CSD).

Advised the President's Office of Science and Technology Policy (OSTP) and federal agencies on the implementation and evaluation of R&D programs.Co-authored report recommending high-impact emerging technologies to receive over $100M in federal research fundingCo-authored reports on national innovation policies of Brazil, Russia, and South Korea, with emphasis on the relationship between higher education practices, government oversight, and innovation outcomes.

Site-directed mutagenesis and functional genetic selections

Worked on the design and execution of a high throughput screen for identifying neuroactive drug compounds based on image analysis of the behavior of zebrafish embryos.