Dominic DeCarlo

University of Pittsburgh

AI & Automation Intern

• —Engineered an ensemble ML model predicting customer experience scores and root-cause risk factors for IT incident tickets, saving managers 98% of manual triage time and projecting ~$110k in operational savings.
• —Assessed frontier model vulnerabilities and developed AI security frameworks to enable the safe deployment of high-performing LLMs.

IT Intern, Data Insights & Reporting

• —Built a Python-based AI/ML model to forecast contribution trends, helping the team proactively strategize for the high-volume giving season.
• —Performed data analysis on complex datasets to develop business-focused data visualizations (Sankey charts, Power BI dashboards), improving reporting efficiency and providing actionable insights into donor behavior.

Camp Counselor

• —Led and supervised daily activities for a diverse group of children, including many with disabilities, from 8 AM to 4 PM, with additional aftercare responsibilities until 6 PM.
• —Fostered an inclusive, supportive environment, ensuring every child had a positive camp experience.

Ground-Up Large Language Model

• —Architected a custom transformer-based neural network from the ground up, executing distributed pre-training pipelines across high-performance A100/H100 GPU clusters to optimize the model architecture.
• —Fine-tuned conversational behavior using a curated proprietary dataset and engineered a stable deployment pipeline to serve the optimized model weights via a live, web-hosted interface.

VIGIL (First Place, Steelhacks XIII)

• —Engineered an automated surveillance system leveraging custom trained video encoders for rapid anomaly screening paired with multimodal VLMs to analyze incidents across continuous CCTV footage in real time.
• —Developed a full-stack operator workspace featuring automated risk scoring, timestamped incident timelines, and multi-channel audio dispatch via speech synthesis APIs.

Fine-Tuned Local AI Coding & Voice Assistant

• —Fine-tuned a 9B parameter LLM via QLoRA to optimize code generation and reasoning, surpassing the larger Qwen 3.5 27B model on 7 standard coding benchmarks.
• —Deployed a fast, local voice agent using STT & TTS pipelines with an LM Studio backend for live hands-free use.

Real-Time Prediction Market Arbitrage SaaS

• —Engineered a Python scanner to instantly detect profitable arbitrage bets across Kalshi and Polymarket.
• —Built an interactive dashboard to sort live market spreads by profit margin with direct execution links to exchanges.

AI Resume Builder for Veterans

• —Built an AI resume builder with Next.js that translates military roles into professional, job-ready bullet points.
• —Designed a complete workflow for users to generate, customize, and download a polished resume online.

College Real Estate Predictor

• —Created a machine-learning model to provide more accurate current and future housing value and rent predictions than Zillow estimates, tailored for landlords and investors.
• —Merged and gathered student and property data using Pandas for analysis and predictions.

AI Soccer Analysis System

• —Developed an AI/ML system using Python and computer vision to analyze football/soccer matches.
• —Implemented real-time object detection to analyze player movement and important game events.
• —Provided valuable insights to improve match evaluation and strategy.

AI Affirmations Generator

• —Built a web application that generates personalized affirmations using AI.
• —Implemented a design-focused, user-friendly interface for users to provide input via chatbox and receive tailored affirmations.

Live Translator

• —Built a real-time web app that translates live speech across languages using Flask and speech recognition.
• —Integrated speech recognition to capture spoken words and Google Translate API for multilingual translation.

Pitt Computer Science Club

• —Contributed to collaborative projects and technical discussions, sharpening problem-solving and technical knowledge.
• —Participated in coding competitions and workshops, deepening fluency in languages and tooling.

Panther Equity

• —Gained hands-on finance experience through market trend analysis and equity research for simulated portfolios.

Python, TypeScript, JavaScript, Java, C, C++, Go, SQL, R, HTML, CSS, Bash, MIPS Assembly

PyTorch, TensorFlow, Scikit-Learn, NumPy, React, Next.js, Svelte, Node.js, Flask, OpenCV, YOLO, LangChain, Tailwind

GitHub, Git, Jira, n8n, Power BI, VS Code, Excel, Markdown