Adi Parekh

Incoming UC Berkeley · Data Science

Adi Parekh

Hi, my name is Adi and I am entering UC Berkeley this fall as a Data Science student. I enjoy building tools to solve all sorts of problems I come across. My website displays examples of this through some of the work I have done.

UC Berkeley Data Science, Fall 2026
4.0 · 4.85 Unweighted / Weighted GPA
Ranked 1st Inderkum High School
3x National C-SPAN StudentCam

Work

Lalitar

Founder

Python Machine Learning Real Estate Data Analysis

AI platform that predicts commercial real estate futures 6–18 months out.

  • Forecasts property sales, refinances, defaults, tenant loss by combining loan data, public records, foot traffic, permits.
  • Most tools show what a building was. Lalitar shows what it's about to become.
  • Turns raw signals into actionable predictions for investors, lenders, and operators.
Visit Lalitar

Sacramento Municipal Utility District

Software Engineering Intern

Summer 2025

Python HTML Data Visualization

1 of 12 interns from 250+ applicants. Shaped EV charging strategy for 600k smart meters across Sacramento metro.

  • Analyzed demographics, transport, energy data to identify neighborhoods for EV infrastructure expansion.
  • Validated geospatial GIS data and built dashboards surfacing adoption trends to engineering teams.
  • Work directly fed infrastructure investment decisions.

CloverLeaf Solutions

Software Engineering Intern

2024

Data Pipelines QA HTML
  • Built SQL-based data extraction and validation pipelines for enterprise clients.
  • Developed dashboards and reports surfacing client trends for business stakeholders.
  • QA and debugging across data pipelines and web tools.

Research

Investigating the Impact of the MACD Technical Indicator on Linear Regression Models for Predicting Financial Time Series

Independent research paper · 25 pages

Does adding the MACD technical indicator to a linear regression model justify the added runtime and memory cost?

Benchmarked two models (baseline vs. MACD-enhanced) across diverse tickers (AAPL, BABA, TSLA, SPY). Finding: MACD improved consistency but rarely accuracy. Extra feature costs ~0.02 MB per model—negligible alone, but scales to real expense across 100k+ models at major institutions.

Python, pandas, NumPy, scikit-learn, tracemalloc

Read the full paper

Projects

Invasive Carp Awareness Game

40+ hours of Python to simulate invasive carp population dynamics in my local pond. Solved sprite-collision with real-time Euclidean distance checks.

IB Biology adopted it for their ecosystem unit; City Councilmember Lisa Kaplan featured it in a newsletter reaching thousands. Showed me how computing + education can drive sustainability.

Python

Play the game

SignalForge Stock AI Predictor

Production ML platform predicting next-day stock prices. Auto-downloads 5 years of market data, extracts 15+ technical indicators, trains 4 models (LR, RF, GB, XGBoost), keeps the best.

FastAPI + SQLite backend, 5-page Streamlit dashboard with Plotly charts, fully containerized Docker + GitHub Actions CI.

Python, Machine Learning, FastAPI, Streamlit, Docker

Recognition

2026

C-SPAN StudentCam · 2nd Prize, Western

"Life, Liberty, and Left Behind" — veterans' rights documentary.

Watch
2025

C-SPAN StudentCam · 3rd Prize, Western

Sacramento levee infrastructure. Interviewed Mayor and engineers; led to official meetings on resilience.

Watch
2023

C-SPAN StudentCam · 3rd Prize, Western

"The Cost to Survive" — healthcare affordability.

Watch

Certification

PCEP: Certified Entry-Level Python Programmer

OpenEDG Python Institute

Earned through structured coursework + hands-on projects: calculators, sorting algorithms, applied learning over passive reading.

Python fundamentals · data structures · control flow