Chris Dollo

Software Engineer & ML Researcher

Education M.S. Computer Science, UMD

Focus Area Computer Vision & ML Systems

Software engineer and ML researcher with industry and academic experience in computer vision, gesture recognition, and full-stack development. Currently pursuing an M.S. in Computer Science at UMD. Co-author of a 2025 Nature Scientific Reports publication on hand gesture reconstruction using motion synergies.

Resume GitHub

Experience

May 2026 – Aug 2026 Peoria, IL

Data Science Intern

Caterpillar Inc.

Developing data-driven ML models to support predictive maintenance and operational analytics across heavy equipment product lines.
Building and evaluating pipelines for large-scale structured and time-series data using Python and SQL.
Collaborating with cross-functional engineering teams to translate business requirements into scalable analytical solutions.

Python Machine Learning SQL Data Science

May 2025 – Aug 2025 Catonsville, MD

Machine Learning Intern

Verizon

Designed and deployed an end-to-end CV pipeline using YOLOv8 to automate equipment inspection, achieving 94% accuracy across 14,000+ labeled images and eliminating manual review for 3 product lines.
Preprocessed and augmented datasets from Roboflow and Hugging Face — improving model robustness and reducing false detections by 20% in production.
Optimized PyTorch training and inference workflows, cutting average latency by 35% and enabling real-time deployment on edge hardware.
Built automated unit and integration test suites, validating 100% of production inference endpoints before deployment.

YOLOv8 PyTorch Python Computer Vision

Jan 2024 – Present Catonsville, MD

Research Assistant

UMBC Sensorimotor Control Lab (Dr. Ramana Vinjamuri)

Architected a real-time Python gesture analysis pipeline using OpenCV and MediaPipe, processing 50+ gesture classes to extract biomechanical motion primitives.
Engineered Gaussian-based joint angular velocity models achieving 95.78% reconstruction accuracy for natural grasps and 92.99% for mudras — outperforming baselines by up to 10.48%.
Designed EMG-based ML classifiers (SVM, LSTM, Random Forest) for prosthetic gesture recognition across multi-channel sensor datasets.
Co-authored a Nature Scientific Reports publication (2025) as 2nd author.

Python OpenCV MediaPipe Deep Learning EMG

Projects

Hackathon

Project Falcon

YOLOv8 CV pipeline to automate cell tower equipment inspection — 94% accuracy across 14,000+ labeled images, reducing false detections by 20%.

YOLOv8 · PyTorch · Python · Roboflow

EcoRecyclr

1st place at the Booz Allen Hamilton Code for Good Hackathon. Android app for real-time waste classification using ResNet-18 fine-tuned on 10,000+ images — 95% accuracy.

PyTorch · ResNet-18 · Android (Kotlin) · Flask

GitHub Demo

Research

Gesture Recognition

Real-time system identifying 32 distinct hand gestures live from webcam using OpenCV and MediaPipe.

Python · OpenCV · MediaPipe

GitHub Demo

PutEMG Classifier

Deep learning pipeline classifying 24 hand movements from 24-channel EMG sensor data — preprocessing, feature extraction, and model evaluation.

Python · PyTorch · MATLAB

Games

Pacman

Browser-based Pacman clone built from scratch — ghost AI, score tracking, and classic arcade mechanics.

HTML · JavaScript

GitHub Play

Publications

Reconstructing hand gestures with synergies extracted from dance movements

Parthan Olikkal, Chris Dollo, Akshara Ajendla, Ann Sofie Clemmensen & Ramana Vinjamuri

Scientific Reports • 2025

A novel framework for understanding hand gestures through movement primitives derived from mudras — hand gestures from Bharatanatyam. Synergies extracted via Gaussian-modeled joint angular velocities reconstruct 75 hand gestures, achieving 95.78% accuracy for natural grasps and 92.99% for mudras, outperforming baseline models.

DOI