Leonardo CortezGentile

Baseball analyst and international scout. I build models that translate pitchers across leagues and data products that help the game reach its fans.

Williams College ’26, Statistics & Computer Science

Résumé LinkedIn @Pitching606 Read the reports
Portrait of Leo Gentile in a navy blazer and plaid tie
About

A lifelong student of the game

I grew up in New York going to ball games and playing in Little League with my dad. Baseball is my family’s biggest tradition, and I fell in love with baseball cards at a very young age. I loved trading stories with my family, hearing about the players of their day and telling them about the ones I follow now. These conversations started a lifelong love of learning, numbers, and storytelling. By the end of Little League, pitching had become my obsession, both as a science and as an art form: how a pitching staff is built, how pitchers are evaluated, and how they change when they move between leagues.

That curiosity followed me to Williams College, where I studied Statistics and Computer Science and spent two summers abroad on research fellowships, scouting KBO pitchers in South Korea and then NPB pitchers in Japan. This work has become an R Shiny app that projects how pitchers from KBO, NPB and CPBL translate to MLB.

In summer 2026, I joined MLB’s Fan Data & Insights team, where I rebuilt forecasting models and developed an AI player tagging system used across the league office. This fall I’m moving to the field side, starting with Hawk-Eye replay and medical systems at Rutgers football. As my career grows, I want to keep learning, finding the stories inside the numbers, and growing baseball globally.

Experience

Career log

  1. College Sports Systems Technician Current
    Hawk-Eye Innovations · Rutgers Scarlet Knights
    Sep 2026 – Present · 2 mos · New Brunswick, NJ
    • Completed onboarding and attended team meetings ahead of live Big Ten football games for the Rutgers Scarlet Knights.
    • Will support the setup, testing and operation of Hawk-Eye SMART replay and medical systems on game days.
    • Will run pre-game system checks and troubleshoot hardware, software and signal-flow issues to keep the technology reliable.
    • Will coordinate with officials, broadcasters and other Hawk-Eye operators to monitor systems and resolve issues in high-pressure live environments.
    • Will install, test and deploy new Hawk-Eye systems and products for Division I football operations.
    Replay systemsLive operationsSports technology
  2. Fan Data & Insights Intern
    Major League Baseball · League Office
    Jun 2026 – Aug 2026 · 3 mos · New York, NY
    • Designed a RAG system that uses rosters, jersey numbers, headshots, social media captions and fuzzy matching to improve the accuracy of an in-production, Gemini-based player-tagging model.
    • Overhauled the MLB.TV single-game viewership forecast with XGBoost and higher-fidelity features, including rolling playoff odds, viewership, unique users and subscriber counts.
    • Developed a “Free Game of the Day” ranking algorithm to inform programming and ad-placement decisions.
    • Engineered an automated “reasons” feature that explained 354 large attendance-forecast misses and flagged teams with systematic forecasting bias.
    • Built a centralized royalty reporting dashboard analyzing transaction data across all 120 Minor League clubs to surface growth trends and top-performing categories.
    • Presented to the SVP of Business Analytics & Operations and the SVP of Business Development, and was recognized for work that laid the foundation for future predictive models and analytics efforts.
    XGBoostBigQueryGeminiLooker
    See the projects →
  3. Baseball Operations Roundtable
    Milwaukee Brewers
    Jun 2026
    • Presented my independent study on count effects, pitch tunneling and sequencing, with case studies on the Marlins and Yoshinobu Yamamoto.
    See the roundtable →
  4. Independent Study, Pitch Analytics
    Williams College Statistics Department · Advisor: Prof. Richard De Veaux
    Feb 2026 – May 2026 · 4 mos · Williamstown, MA
    • Modeled the effects of count, pitch sequencing and pitch tunneling on the full 2025 MLB Statcast dataset.
    • Identified optimal pitch-usage adjustments for Marlins pitchers after the club changed its pitch-calling approach.
    • Concluded that tunneling works best when pitches with tighter separation at the tunnel point need less separation at the plate to be effective.
    • Compared neural networks to first-order Markov models for whiff prediction and found that they marginally improve predictive capability.
    Read the 82-page paper →
  5. Research Fellow, NPB Scouting
    Wilmers Memorial Student Travel Fellowship
    Jun 2025 – Sep 2025 · 4 mos · Tokyo, Japan
    • Conducted a six-week sabermetric case study of 1,300+ NPB pitchers, applying multiple linear regression to predict Wins Above Replacement (WAR).
    • Developed regression models and applied stepwise and interaction selection to identify the strongest predictors of NPB pitcher WAR.
    • Implemented and built on the indicators most predictive of NPB WAR and future MLB success (K-BB%, GB%, SwStr%, xFIP, WPA).
    • Consulted with six MLB, MLB Japan and NPB front-office analysts on global scouting strategy and the use of Hawk-Eye data in pitcher evaluation.
    • Generated scouting reports on 22 top NPB pitchers using DeltaGraphs, the 2023 World Baseball Classic, the 2024 Premier12 and Yakyu Cosmopolitan.
    South Korea & Japan work →
  6. Research Fellow, KBO Scouting
    Class of 1945 World Fellowship Program
    Jun 2024 – Sep 2024 · 4 mos · Seoul, South Korea
    • Coordinated in-depth scouting and statistical analysis of 164 KBO pitchers across 43 advanced metrics (WAR, FIP, SIERA and more) to assess their potential for MLB transitions.
    • Refined predictive models of player performance with stepwise, ANOVA, VIF and transformation-based approaches.
    • Sharpened my evaluation across the KBO and MLB talent pools and developed player development ideas with a KIA Tigers analyst in Gwangju.
  7. B.A. Statistics & Computer Science
    Williams College
    Aug 2022 – Jun 2026 · 4 yrs · Williamstown, MA
    • Dean’s List. Captain, Williams Ultimate Frisbee (’26). Benedict Prize in Ancient Greek (2024).
  8. Front Facing Marketing Intern
    The Sideline Hustle · Football education and training
    Jun 2020 – Jun 2024 · 4 yrs 1 mo · Atlanta, GA
    • Worked with 40+ NFL and NCAA football players to maintain and grow the company’s client base and outreach.
    • Organized and updated a large dataset of player information used to build new training materials and workouts.
    • Created marketing tools and content for “Routes like Picasso,” an interactive IGTV series that drew 50,000+ unique impressions per post and reached 100,000+ users.
    MarketingContentAthlete relations
International scouting

Two summers, two leagues

I spent summer 2024 in South Korea and summer 2025 in Japan watching games in person and building models to find which pitchers are ready for MLB.

Sajik Baseball Stadium in Busan at sunset, packed stands under the mountains
Sajik Baseball Stadium, Busan, home of the Lotte Giants. Photo by Leo, summer 2024.
South Korea · KBOSummer 2024

Kimchi and Curveballs

A six-week scouting trip and statistical study of KBO pitching, built to find which arms could make the jump to MLB.

Scatterplot of actual vs. predicted WAR for KBO pitchers, with Kyle Hart, Aaron Wilkerson, Ariel Jurado, Charlie Barnes and Gwak Been labeled near the top
Actual vs. predicted WAR for every qualified KBO pitcher, 2024. Points near the diagonal were predicted well.
The modelBackward stepwise · adj. R² = 0.879

WAR = 0.878 − 0.314·GamesRelieved + 0.229·Hits − 0.052·HomeRuns + 0.054·Walks + 0.309·Strikeouts − 0.358·ERA − 0.264·WHIP + 0.133·HR/9 + 0.144·OBPAgainst − 0.204·OPSAgainst

  • Scouted games in person across South Korea, from Seoul to Sajik Stadium in Busan.
  • Analyzed 164 KBO pitchers across 43 advanced metrics (WAR, FIP, SIERA and more) for MLB-transition potential.
  • Worked with a KIA Tigers analyst in Gwangju on how clubs evaluate pitchers.
  • Built a regression model of KBO pitcher WAR, then narrowed its top 20 by pitch run value, velocity and age.

Top 5 MLB prospects in the KBO (2024)

1An Woo-jin
2Ariel Jurado
3Enmanuel De Jesus
4James Naile
5Kyle Hart
KBO report (43 pp) →
Japan · NPBSummer 2025

Sushi and Sliders

A six-week scouting trip and statistical study of five seasons of NPB pitching, built to find the league’s most MLB-ready arms.

Scatterplot of predicted vs. actual WAR for the NPB test set, tightly clustered around the fit line
Predicted vs. actual WAR on the held-out 2025 NPB test set (r = 0.955).
The modelStepwise + interactions · adj. R² = 0.842

WAR = −0.12 + 0.014·StartingPitcherIP − 0.202·trueRunsAllowed + 0.010·RE24 − 0.059·Homeruns + 0.016·RelieverIP + 0.014·Strikeouts − 0.017·Walks − 0.010·Hits + 0.020·IntentionalWalks + 0.041·GamesStarted + 0.004·FIP− + 1.198·BAA − 0.101·WPA − 0.009·StolenBases − 0.021·Losses + 0.074·pLI + 0.000048·(StartingPitcherIP × Strikeouts) − 0.001·(Homeruns × Walks)

  • Scouted games in person in Japan and wrote reports on 22 top NPB pitchers.
  • Consulted with six MLB, MLB Japan and NPB front-office analysts on global scouting strategy and Hawk-Eye data.
  • Studied 1,300+ NPB pitchers and built a regression model of NPB pitcher WAR, then layered in WBC, Premier12 and DeltaGraphs data.

Top 5 MLB-ready NPB pitchers (2025)

1Tatsuya Imai
2Jon Duplantier
3Hiromi Itoh
4Foster Griffin
5Hiroto Saiki
NPB report (20 pp) →
Summer 2026 · Major League Baseball

Predictive models & business insights

A summer on the Fan Data & Insights team. Four projects, each presented to MLB’s senior vice presidents of Business Analytics & Operations and Business Development.

AI player tagging · RAG
RAG + Gemini

Fixed a caption-ingestion bug in the production tagging model, tightened the prompt, fed daily rosters and jersey numbers in as live context, and added fuzzy matching for typos, accents and suffixes. The model tags which players appear in MLB social posts for reporting and content strategy.

MLB.TV viewership forecast
XGBoost

Rebuilt the single-game viewership model with decay-weighted head-to-head history, daily subscriber counts built from fan avidity, and rolling playoff odds. It powers the Free Game of the Day pick and ad placement.

Attendance error analysis
354big forecast misses explained

Sorted every large attendance-forecast miss into causes (scheduling, matchup, team performance, weather, local events, unlogged promotions) and flagged clubs the model consistently over- or under-predicts.

MiLB e-commerce royalties
120Minor League clubs

A Looker dashboard tracking online merchandise sales by club and product category, built for MiLB club leadership and commercial teams to spot growth trends and top sellers.

Fall 2026 · Hawk-Eye Innovations

On the field side of the data

Hawk-Eye cameras feed Statcast in every MLB park. At Rutgers I’ll be supporting the same company’s replay and medical systems on Big Ten Saturdays, a firsthand look at how tracking data is captured before it ever reaches an analyst.

  • Completed onboarding and attended team meetings ahead of live Big Ten games.
  • Will support the setup, testing and operation of SMART replay and medical review systems on game days.
  • Will run pre-game checks on hardware, software and signal flow, and troubleshoot live when something drops.
  • Will work alongside officials, broadcasters and other Hawk-Eye operators in a high-pressure live environment.
  • Will install, test and deploy new Hawk-Eye products for Division I football operations.
Featured project · R Shiny

International Pitcher Translator

When a pitcher moves from the KBO, NPB or CPBL to MLB, what should a club expect? This app answers it with seven projection models, career odds, and real historical comps.

Inside the app

  1. Scouting Report
  2. Next Up
  3. Explorer
  4. Profile
  5. Quadrants
  6. Translations
  7. Survival
  8. Accuracy
  9. Article

Data behind it

LeagueSeasonsPitcher-seasons
KBO1982–20257,662
NPB1936–202518,284
CPBL1990–20253,308
MLB1871–202549,094
US minors (AAA/AA)1996–2025304
  • Two rules of thumb plus mixed, Bayesian, random forest, boosted-tree and two-way league models, each scored on pitchers it never saw. The best misses first-year FIP by 0.61 runs; luck alone accounts for 0.45.
  • New: Taiwan’s CPBL, projected by the two-way league model, which learns from 7,040 seasons by 1,103 pitchers who pitched in two or more leagues.
  • Career odds for KBO and NPB pitchers: the chance he reaches MLB and sticks two or more seasons.
  • KNN/PCA comps matched within role, plus Kaplan-Meier survival curves.
Launch the app →Read the app guide (14 pp) →
Scouting Report tab projecting Yoshinobu Yamamoto's 2023 NPB season to MLB year one: a chart and table comparing the mixed, Bayesian and two-way league models
Scouting Report: Yoshinobu Yamamoto, NPB 2023 → MLB year 1, model by model.
Comp map placing Yamamoto's 2023 NPB season among MLB seasons near Jake Arrieta 2015 and Zack Greinke 2009, beside his closest past movers Masahiro Tanaka, Nick Martinez and Brandon Waddell
Comps: closest MLB match is Jake Arrieta, 2015; closest past mover is Masahiro Tanaka.
Writing

Scouting notebook

Scouting reports, player plans and takes on prospects, pitches and roster moves. Open any post to read it in full.

Portfolio

Reports & research

My research papers, scouting reports and slides, in full. Each opens as a PDF in a new tab.

Milwaukee Brewers · Baseball Operations Roundtable · June 2026

Brewers Roundtable

Brewers Baseball Operations Roundtable graphic reading I'm attending

I’m excited to share that I was selected to present at the Milwaukee Brewers’ Baseball Operations Roundtable this June.

I presented my senior independent study, An Analysis of 2025 MLB Statcast Data: Count Effects, Tunneling, and Pitch Sequencing, to members of the Brewers’ front office. Thank you to the Brewers for the opportunity, and to Professor Richard De Veaux for advising the project.

What I presented

  1. Marlins pitch usage before and after their pitch-calling change
  2. Ryan Weathers’ 2026 profile
  3. League-wide tunnel distance vs. whiff rate
  4. Yoshinobu Yamamoto: elite without elite tunneling
See the slides → Read the full paper →
My hobby

How I kill time

Spin for one of the 30 MLB organizations, spin again for one of its four full-season affiliates, then pull up that club’s games and write a scouting report on someone on the roster. Try it.

Spin 1 · OrganizationNot spun yet
Spin 2 · AffiliateWaiting on the club spin
From the stands

Photos & video

Ballparks in South Korea, Japan and the U.S., two halls of fame, Monument Park, and a few from growing up. Tap any photo to see it larger.