CTO, Incite.ag · Boston, MA
Integrated science,
agentic systems.
I build applied machine learning and agentic systems — and I got here by refusing to pick a department. The interesting problems sit in the seams between fields, where one discipline's open question is another's solved commodity.
- Currently
- Chief Technology Officer, Incite.ag
- Background
- Neuroscience & electrophysiology
- Working on
- Knowledge codification for machine reasoning
The premise
A single brain cannot hold the coupling from gluons to atoms to neurotransmitters to cells to populations. Computation is the only tool I know of that holds a structure larger than the head that built it — provided the knowledge is codified well enough to reason over. That proviso is most of the job.
What I do
Engagements →- 01
Owning a platform end to end
Application, data infrastructure, cloud and hybrid systems, integrations, operations, and security — as one system rather than six handoffs. I still spend most of my week in the code, which is the only way I know to keep architecture honest.
- 02
Document intelligence and extraction
Agent-driven pipelines applying OCR, NER, and structured extraction to messy real-world corpora — scanned records, semi-structured PDFs — and mapping them into schemas you can actually query and audit.
- 03
Regulated, auditable data systems
Translating regulatory guidance and scientific literature into domain models that hold up under audit. When the rules change mid-quarter, the pipeline has to change with them without losing traceability of what was reported and why.
- 04
Applied ML that survives real data
Forecasting, computer vision, and evaluation on proprietary data that is partial, mislabeled, and disagrees with its own documentation. Most of the work is the pipeline and the error analysis, not the architecture.
Selected projects
All work →Legacy systems, agents, and the human on the hook
Extraction pipelines applying OCR, NER, and structured extraction to a corpus exceeding 10,000 pages annually, feeding regulated reporting — and the engineering discipline that decides where agents are allowed to touch it.
Read the writeup →Deep neural network interpretability against a ground truth
A novel method for assessing multi-way feature explainability in deep networks, evaluated against a new ground truth measure rather than against other explanation methods. First author, AAAI 2023 R2HCAI workshop.
Read the writeup →MVPA on EEG from tri-polar concentric ring electrodes
Applied multivariate pattern analysis to EEG to classify event-related potentials, and to characterize what information conventional electrodes capture versus tri-polar concentric ring electrodes. Published as a VSS abstract in Journal of Vision.
Read the writeup →
Latest writing
All writing →Why I went looking for a major that didn't exist
The silos between the natural sciences are an administrative fact, not a fact about the world. On arguing with department chairs, recording evoked potentials, and why the ceiling on what one brain can hold pushed me toward computation.
ELIZA is still the most important result in AI
Surface fluency carries almost no information about the mechanism underneath. That was true of a string-matching program in 1966 and it's true of everything shipping today, except now we've lost the ability to measure the gap.
Have a problem in the seams?
The work I like best is the kind where the technique already exists somewhere and nobody has carried it across yet. If that sounds like what you're sitting on, get in touch.