Open to Data Scientist & ML Engineer roles
Tenicka Norwood.
Physicist turned educator turned data scientist. I've managed 36+ people, shipped a library to PyPI, and published 5 IEEE papers. Right now I'm figuring out how to make federated learning fair without sacrificing privacy.
How I Got Here
The short version: I keep finding new ways to use the same instinct. See a system, find the pattern in the data, make the model better. The tools change. The thinking doesn't.
The Physicist
Started as a physics teacher. Learned that the universe is model-dependent, and that models get better over time if you feed them better data. That idea never left.
Where I've Worked
Click a role to dig in. The star gives you the full story.
Conducting research on privacy-preserving federated learning, swarm intelligence, and trustworthy AI in healthcare under Dr. Uttam Ghosh.
- Developing SwarmClinical: PSO-based orchestration for non-IID federated learning
- Architecting multimodal Digital Twin fusing time-series with clinical narratives
- Published FairSwarm library on PyPI for fair coalition selection
- 3 accepted papers at IEEE SoutheastCon 2026
What I'm Good At (and What I'm Not)
I'd rather you know upfront where my gaps are than find out after the interview.
Strong
- Python:7+ years, primary language across all roles
- PyTorch:FairSwarm, neuro-symbolic, graph anomaly detection
- Power BI:semantic models, DAX, federal client dashboards
- Federated Learning:FairSwarm on PyPI, SwarmClinical, Flower
- SQL / ETL:pipelines at NYC DOE, Excella, BigQuery for MIMIC
- Team Leadership:36+ people managed across two organizations
Moderate
- TensorFlow / Keras:CNNs for medical imaging
- D3.js:won Iron Viz, not daily driver
- MongoDB:EduCore platform backend
Gaps (I'll Tell You)
- Snowflake:BigQuery is her cloud warehouse
- Spark / PySpark:limited production use
- Kubernetes:no orchestration experience
Am I Right for Your Role?
Paste a job description. I'll tell you where I match, where I don't, and whether it's worth a conversation.
Job descriptions are not stored. Analysis is ephemeral and runs per-request.