Draft. Not finished, and visible so it cannot be forgotten.
How I work
Replace this lede with two or three sentences on what you actually believe about building software. Not what sounds good in an interview — what you would still say after a bad week.
Replace this with a principle
replace this with where it came from
Write the principle as a sentence someone could disagree with. “Write good code” is not one; “I would rather ship the unglamorous half first and be boring about it” is. Then spend a paragraph on what it cost you the last time you held to it.
The em-dash line above is the anchor and it is the point of this whole page. Name the real job, number or decision: PNNL, 300,000 mass-spec files, or MatX, arguing about what efficiency means, or Datathon, 300 patient scenarios. If a principle has no anchor, it is a slogan, and the page will render a visible gap where the anchor should be until you write one.
Replace this with a second principle
replace this with where it came from
Three or four principles is a page. Seven is a manifesto and nobody finishes it.
Keep each one to a paragraph or two. If one of them needs a list, you can write bullets:
Bold, italic, backticked code and links all work the way they do in Markdown. Straight quotes and apostrophes are converted to real ones on build, so type normally.
WORKFLOW
- Edit this file. Save it.
npm run dev reloads the page on its own. - When it says what you mean, tell Claude to format it — that is when the typesetting, the trimming and the final structure get done.
- Set
Draft: to no at the top. The draft band disappears and the window stops calling itself a draft.
Everything below the WORKFLOW line is part of the document too, so delete this section when you start writing for real.
Jayden Lee
I'm a Data Science and Computer Science student at UC Berkeley who
builds machine learning systems for data that doesn't behave:
behavioral logs, mass spectrometry output, clinical time series.
Most of my work has been the unglamorous half of ML: getting 300,000
files through a pipeline without it falling over, deciding what a
false discovery rate should be, arguing with hardware engineers about
what "efficiency" means. I like problems where the modeling is only
interesting once the data engineering is honest.
- Study
- B.A. Data Science and B.A. Computer Science, UC Berkeley
- GPA
- 3.63
- Group
- Open Project · Sports Analytics Group at Berkeley
1st place · 7th Annual Datathon at Berkeley
Clinical Intervention Simulator
A Monte Carlo framework that models 300 patient deterioration
scenarios, feeding a Random Forest that optimizes interventions in
real time.
Screenshot · 21:9
The simulation is vectorized with NumPy so a full sweep runs in
seconds rather than minutes, which is what made it usable as an
interactive tool instead of a batch job. It ships as a Streamlit app
over a Pandas ETL pipeline that reconciles multimodal health data into
one frame.
Built at the 7th Annual Datathon at Berkeley, where it took first
place against more than 50 teams and over 200 participants.
- Stack
- Python · scikit-learn · NumPy · Pandas · Streamlit · Tableau
- Scale
- 300 patient scenarios
ChessBlitz
An open-source teaching platform in use at Berkeley Chess School.
Screenshot · 16:10
Flask and Supabase behind the storage layer, OpenRouter generating
hints tailored to the position a student is actually stuck on, and a
React dashboard where a teacher runs their own classroom: rosters,
assignments, and who is struggling with what.
- Stack
- Flask · Supabase · OpenRouter · React · TypeScript · SQL
- Status
- Open source, in classroom use
Experience
-
Feb 2026 – now
Machine Learning Researcher
Cal Athletics · contract
Clustering and supervised models that predict ticket-purchase
propensity from behavioral data, plus multivariate regression
behind a dynamic fan scoring system. Outputs deploy into CRM
workflows for segment-level personalization.
-
Dec 2025 – now
Machine Learning Consultant
MatX · contract
Leading eight engineers building a competitive benchmarking
platform for a Series A AI-chip startup, evaluating LLM training
efficiency across participant-submitted models. Docker
containerization and Kubernetes job scheduling isolate each
participant. Defined decode efficiency metrics directly with MatX
hardware engineers.
-
Sep – Dec 2025
Business Analyst
Oakland Roots Sports Club · contract
Built a fan-engagement scoring system in Python that folded
real-time BART transit data into a marketing dashboard.
-
Jun 2024 – May 2025
Modeling & Simulation Intern
Pacific Northwest National Laboratory
A target-decoy statistical framework modeling false discovery
rates for peptide identification, with a precision-recall
threshold optimization pipeline. Ran GPU-accelerated proteomics
workflows on the Tahoma HPC cluster under Slurm.
-
Jun – Aug 2023
Computational Biology Intern
Pacific Northwest National Laboratory
Pushed 300,000+ mass-spec files through a
GPU-accelerated Casanovo inference pipeline, and benchmarked
18 bacterial metaproteomics datasets with
Pandas and Seaborn.
B.A. Data Science, UC Berkeley · GPA 3.7
Sports Analytics Group at Berkeley
- Kind
- PDF document
- Size
- 181 KB
- Pages
- 1
Download
spots.map
A few blocks of Berkeley, and what I ate on them. Walk it and the
places open one at a time.
Draft. Not finished, and visible so it cannot be forgotten.
Places I actually went
Twelve or so places in one Berkeley neighborhood, and what I ate at each. Walk the map and they open one at a time.
No places written up yet — the map below shows the storefronts standing empty, which is the truth of it.
Every place on this map is somewhere I actually went, written up in
one file I edit by hand. Nothing here is generated and there are no
filler entries — where I have not written anything, the storefront
stands empty and the map says so rather than inventing a restaurant
to fill the gap. That constraint is most of why it is worth walking.
It is a map rather than a list because a list gets skimmed. Twelve
names and twelve one-line reviews are four seconds of scrolling and
nothing retained. Making you walk two blocks to the next one buys
each place the few seconds it takes to actually read it, which is the
only thing I wanted from the format.
There is no score, no timer and no way to lose, and it does not ask
you to come back tomorrow. Walking is one keypress per block, the map
redraws only when you find something, and nothing on this page
animates while you are standing still.