Rishi Vemulapalli

Engineer.
Builder. ML Tinkerer.

Computer Engineering + Data Science · AI/ML @ UIUC · Graduating May 2027

Open to Summer 2026 internships

See My Work ↓
01 —

About

I build systems that learn, navigate, and ship. From fine-tuning YOLOv8 on embedded hardware to implementing Pure Pursuit path planning for a VEX World Championship robot, I work across the full stack — ML engineering, robotics, and C++ systems. Currently at the UIUC makerspace and actively seeking internship roles in ML, embedded systems, or software engineering.

PythonC/C++JavaJavaScriptTypeScriptSwiftC#PyTorchTensorFlowOpenCVYOLOv8ONNXbi-LSTMKalman FilterPure PursuitPID ControlOdometryNext.jsReactMySQLGitLinuxJupyterFusion 360Unreal EngineMATLABEmbedded MLPythonC/C++JavaJavaScriptTypeScriptSwiftC#PyTorchTensorFlowOpenCVYOLOv8ONNXbi-LSTMKalman FilterPure PursuitPID ControlOdometryNext.jsReactMySQLGitLinuxJupyterFusion 360Unreal EngineMATLABEmbedded ML
02 —

Projects

NLP / Deep Learning

BiLSTM COVID-19 Tweet Classifier

PyTorchbi-LSTMGRUBERTNLP

851k-parameter bidirectional LSTM classifying COVID-19 tweets as informative with 76.8% test accuracy. Custom 9577-token vocabulary with regex tokenizer; standardized padding tripled data-loading speed.

embed_dim=75 · hidden_dim=96 · dropout=0.6

76.8% test accuracy · 851k params

Data Science / Statistics

Formula 1 ML Pit Stop Analysis

PythonPandasScikit-learnRegressionEDAStatistics

Analyzed F1 pit stop data (1950–2024, 21k+ rows) across 35 circuits and 76 drivers. One-sample t-tests showed Mercedes, Ferrari, and Red Bull are statistically significantly faster (p < 0.01). Linear regression using constructor identity and race features.

R² = 0.32 · RMSE ~2.8s · 21,184 rows

Robotics / Algorithms

VEX Autonomous Navigation Stack

C++Kalman FilterPure PursuitPIDOdometryPROS

Full autonomous navigation suite for Illini VEX Robotics — 3-state Kalman filter fusing dual IMUs and accelerometers, differential drive odometry, async PID with integral clamping, and Pure Pursuit path-following with circle-intersection lookahead geometry.

2024 VEXU World Championship qualifier · top 30% globally

ML / Sensors

ML Fitness Tracker

PythonMatplotlibSignal ProcessingSensor FusionML

Real-time activity recognition pipeline using sensor data visualization and processing. ML models predict user activities and track exercise rep counts. Applies outlier detection, signal filtering, and dimensionality reduction to maximize accuracy.

Real-time inference · multi-activity classification

Systems / C++

Stock Market Tracker

C++libcurlImGuiCSV ParsingDockerSSL

C++ application fetching real-time stock data from Yahoo Finance via libcurl with SSL. Custom candlestick chart rendering by extending the ImGui plot library. Docker-based dev environment with Catch2 unit tests for URL generation and data retrieval.

Custom candlestick renderer · real-time OHLC data

03 —

Experience

Siebel Center for Design — The Shop, UIUC

Shop Attendant · Aug 2025 – Present

Training students on laser cutters, CNC, and waterjet. End-to-end fabrication workflows from design prep to prototyping.

Illini VEX Robotics, UIUC

Programming Lead & Full-Stack Developer · Aug 2023 – Present

Designed robot localization and path planning algorithms. Drove 20% efficiency increase via parallel processing. 2024 VEXU World Championship qualifier. Also building the team's responsive web platform.

RoboHawks Robotics Club Inc. — Jericho, NY

Co-Founder, Lead Programmer, Mentor · Sep 2018 – Present

Co-founded a FIRST Tech Challenge non-profit at age 16. Led mechanical, electrical, and programming. TensorFlow computer vision for robot autonomy.

Jericho VEX Robotics Club — Jericho, NY

Captain, Lead Programmer · Jun 2022 – Jun 2023

Competed at the 2023 VEX Robotics World Championship. Implemented PID, Odometry, and Pure Pursuit control systems on competition robot.

04 —

Let's Talk.

Open to Summer 2026 internships — ML engineering, embedded systems, or software engineering.