High school junior in Newton, Massachusetts. I help a startup's CEO size its market, audit the LiDAR scans that train its AI, and research college closures with a professor in Python. Four sports, one podcast, Class of 2028.
Liam Kranenburg
From D.C. to Newton
Born in Washington, D.C. in 2009. My family moved to Newton, Massachusetts in 2018, when I was in 2nd grade.
Born in Washington, D.C.
Moved to Newton, Massachusetts
Newton has been home ever since.
Started at Newton South
- Elected class president for my grade
- Played soccer in the fall
- Started 16 months of weekend shifts as a busser at Marseille in Boston
Four sports and a podcast
- Alpine ski racing in the winter, then baseball in the spring and golf
- Swim aide at Camp Billings for the summer
- Started The Kranenburg Show with a friend in September
The year I started building
- Became president of the Senior Smiles Club (March) and started caddying at Brae Burn (April)
- Started as a student researcher with a University of Houston-Downtown professor (June)
- Counselor-in-training at Camp Billings (June–July)
- Market research intern, then LiDAR data auditor, at a stealth-mode AI startup
- Founded the AI for All Club and started Wovana Web Design
- Built an agent that sent 538 internship applications, and built The Call
- Our podcast earned an honorable mention in NPR's Student Podcast Challenge
Summer internships
Available full-time from June 17, 2027, for AI and tech, finance, business strategy and data roles.
Graduate from Newton South
A dot map of the contiguous U.S. Pick a year and the light moves between D.C. and Newton.
Work and research
Two roles at a stealth-mode AI startup, research with a university professor, and the jobs that came first.
I help the CEO estimate how big the company's market is: utility vegetation management, the work of keeping trees clear of power lines. I build the estimates from Census figures, industry reports, SEC 10-K filings, Edison Electric Institute member records and public utility commission filings across 100+ investor-owned utilities. They feed company strategy and the CEO's pitches.
It's my first real research job, and I've learned a ton just from having to make my numbers make sense.
All U.S. spending on keeping vegetation clear of utility lines.
The company's name and figures stay confidential.
Three nested shells for TAM, SAM and SOM, cut open so you can see inside.
I audit LiDAR point-cloud scans for accuracy, flagging anomalies and misclassifications so the company's AI trains on clean, human-verified data. I work in the company's own scan-review tools: the quality check that should eventually let the model automate the review itself.
It's also why this page is made of points.
What a LiDAR scan of a power line corridor looks like: scan rings on the ground, poles, cyan conductors, purple trees, and white points where branches get too close. An illustration, not company data.
I help Professor Candace TenBrink research college closures and liberal arts degrees. I write Python scripts that clean and combine federal IPEDS data, then turn it into charts, including one that lines up closures with the events around them.
Getting a messy spreadsheet to actually say something is my favorite part.
Raw rows on the left, and a tidy grid of exactly 2,064 bright points on the right, one per closure.
Caddying, busing and camp taught me service under pressure.
A helix with one bright node for each of these five jobs.
Things I've built
Tools for baseball, markets, my own internship search and studying. Each row rebuilds the points.
I got tired of yelling at managers through my TV, so I built The Call.
Set any game situation and it ranks every move (steal, bunt, pinch-hit, intentional walk, infield in) by how much it changes your team's chance to win. It runs on real 2025 data: all 30 lineups, Statcast sprint speeds and catcher arms, park factors and lefty/righty splits.
A baseball with all 108 double stitches.
I backtested a trend-following strategy on Nasdaq-100 futures that buys pullbacks to a moving average, using 44,765 hourly bars from January 2019 to August 2026.
- Tested 48 parameter combinations with 6-fold walk-forward validation (2 years to fit, 1 year to test). 19 passed my filter.
- Audited the backtest engine line by line for lookahead bias
- Stress-tested trading costs and compared risk overlays. In backtests, a volatility filter cut max drawdown about 20% and shortened it by more than half.
- Built a matching TradingView indicator in Pine Script
Backtests only. No real money traded.
My real results. One pillar per combination, height is its mean walk-forward Sharpe ratio, and the 19 bright pillars passed.
A Python bot that trades an Opening Range Breakout strategy through the Alpaca API. The strategy marks the high and low of the session's opening period and trades when price breaks out of that range.
Paper trading only.
Dual Momentum against buy and hold
A 31-year backtest (1995–2026) of a monthly rule that holds U.S. or international stocks, whichever has the stronger 12-month return, and switches to bonds when neither is positive.
An opening range breakout on a candlestick chart. The band is the opening range and the ring marks the breakout. An illustration, not real trades.
An AI-agent setup for internship season. Claude finds new roles every day and builds tailored resumes. Muse applies one role at a time, and an inbox listener tracks the replies. I only step in when something is blocked.
I then cleaned it up so anyone can set it up for their own search.
538 points, one per application, and 9 bright stars for the invites.
Adaptive AI quiz tool
Serves harder questions after correct answers and easier ones after misses. I built it to help a classmate study.
Radar map app
An interactive radar map of 116+ locations.
A radar sweep with 116 blips.
Things I've started
An energy research brief I publish, a web design company I took to market with cold calls, a podcast with a friend, and an AI receptionist I'm building now.
KPA publishes short decision briefs on energy and markets. The first, Load Bearing, shows how AI data centers ended 14 years of flat US power demand, and why the plants to serve them arrive after the forecasts that justify them.
I research, write, chart and fact-check each brief. They're drafted with AI assistance, and every number is recomputed in Python from public data and checked by hand.
A transmission tower carrying the load.
A web design business for landscaping and home-service companies. I wrote the outreach scripts, set the pricing and built the brand.
None of them closed, but I learned that you have to keep trying and not stop until you reach your goal.
A landscaping company's site in exploded view, every layer pulled apart.
A podcast I make with a friend: honest conversations about things that affect teenagers, like peer pressure and social expectations. For one episode we filed a public records request to get original source material.
A voice waveform.
A voice AI receptionist that answers calls, qualifies leads and books appointments. I'm starting with real estate agents. The first client is my mom, a real estate agent who had been missing calls.
Rings going out from one point, like a phone that finally gets answered.
Clubs and community
I founded one club, run another, coach kids' golf, and have 318 verified community service hours.
A club that teaches students practical, everyday uses of AI through hands-on workshops.
I started it by pitching the idea to my friends, then found a teacher to sponsor it, and it has grown to about 50 members.
50 nodes around one hub.
I run our Senior Smiles Club, which connects students with residents at Adelaide of Benchmark. I set up the visits and go most weeks myself.
A heart.
I help coach younger kids the basics of golf, along with the life skills the program is built around.
Mostly I try to keep it fun and make sure they actually want to come back.
A flag on the green.
My class elected me to represent our grade freshman year. I helped plan dances, fundraisers and school events.
My class year.
- Member of Kids Against Cancer since April 2026, helping with awareness, fundraising and outreach for families affected by cancer
- Legacy Scholars Program
- Courageous Conversations on Race
318 bright points, one per service hour.
Four sports
I've played four sports for Newton South since 2024. Golf and baseball are varsity.
Golf runs through a lot of my life: I play varsity, caddie at Brae Burn, and coach kids at First Tee.
A golf ball on a tee.
I play varsity baseball. Off the field I built The Call, a tool that ranks MLB managerial decisions by win probability.
A baseball, 108 double stitches.
I raced alpine for Newton South in the winter of 2025.
A slalom course, gates alternating colors down the hill.
Boys' soccer, my first season at Newton South.
A soccer ball: 12 pentagons, 20 hexagons.
Newton South
I'm a junior in the Class of 2028. My classes this year, a certification, and the languages I speak.
This year
On my own
Programs
A normal distribution, for AP Stats.
And for computers
One orbit per language. Brighter means more fluent.
In my words
I like figuring out how things work and being useful while I do it. Mostly I just like starting things and sticking with them.
I'm interested in how businesses use data to make decisions. I'm drawn to finance and AI because I believe they are the future and will keep growing, and I want to be part of that.
Rooms I've been in lately
-
Oct 2026
Finance & ChatGPT with Davis Clarke, Tufts UniversityHosted by the OpenAI Student Collective and Tufts Financial Group. I was probably the youngest person there.
Go at everything with energy and confidence, and it gets easier.
-
Oct 2026
AI in Finance Lab, BostonAn afternoon with CFOs and controllers. I talked with the CEO of Hyphenate, which builds AI agents that take real work off finance teams.
Find a real problem and build something good for it.
My face, rebuilt in points from the photo at the top of the page.
Let's talk.
I'm looking for summer 2027 internships and pre-college programs in AI and tech, finance, business strategy and data. Full-time from June 17, 2027, and part-time or remote during the school year.
liamlkranenburg@gmail.comOnline
Toolkit
Also into