Resume

Download PDF

Nicholas Fitzpatrick

nicholasfitzpatrick@berkeley.edu · San Francisco Bay Area, CA · linkedin.com/in/n-fitzpatrick

UC Berkeley IEOR master's graduate (May 2026) with hands-on experience building and deploying GenAI systems, retrieval-augmented architectures, and LLM-based workflows. Focused on roles where AI drives measurable operational and business outcomes.

Education

University of California, Berkeley – College of Engineering Aug 2025 – May 2026
Master of Analytics – Industrial Engineering and Operations Research, Berkeley, CA
  • Coursework: Deep Learning (CNNs, Transformers, LLMs), Machine Learning, Supply Chain Analytics, Optimization, Simulation, Data Engineering, Database Systems
  • IEOR Department scholarship for exceptional merit and character
University of California, Berkeley – College of Environmental Design Jun 2018 – May 2020
Bachelor of Arts in Sustainable Environmental Design – with Honors, Berkeley, CA
  • Focused on geospatial and energy data analysis as applied to architecture and urban planning
  • Lectured as a Student Instructor for UC Berkeley School of Public Health on energy, climate, and governance
Santa Rosa Junior College Aug 2023 – May 2025
A.S. Computer Studies – Web Full-Stack Developer | A.A. Computer Studies – Web and Multimedia, Santa Rosa, CA

Professional Experience

Geopogo – Berkeley, CA Oct 2025 – Present
AI Engineer & Product Lead (Internship)
  • Built and deployed an end-to-end pipeline converting architectural drawings into labeled training sets for a building code compliance model; onboarded external architecture firm partners for live compliance testing and production inference.
  • Integrated Google GenAI capabilities for image generation and generative design workflows via Python and JavaScript; developed frontend features and led UX/UI decisions across the platform.
  • Coordinated a cross-functional team of 10 across engineering and product to integrate the data pipeline and compliance model into the core platform; managed iterative delivery from user feedback through to shipped features.
  • Shipped redesigned landing page and consolidated site navigation, improving user onboarding experience for architecture and design firm clients.
Outlier – Remote Oct 2022 – May 2024
AI Data Specialist (Contract)
  • Applied RLHF and hallucination-control evaluation to LLM outputs across agentic and multi-step reasoning tasks, improving accuracy and safety adherence in production-grade systems.
  • Conducted prompt engineering experiments and performed quality assurance on model responses and peer evaluations.
Cook's Vineyard Management – Sonoma, CA Sept 2020 – Oct 2022
Operations Manager
  • Owned end-to-end operations across 475 acres and 15+ clients, managing dynamic resource allocation, real-time exception handling, and crew routing under seasonal SLAs.
  • Built a spreadsheet-based scheduling and monitoring system tracking crew assignments, equipment utilization, and job completion across concurrent sites; reduced idle time and rework via daily reallocation logic.

Projects

RAID – Retrieval-Augmented Inverse Dynamics for Robotic Manipulation Spring 2026
INDENG 242B – Deep Learning
  • Designed a retrieval-augmented architecture solving inverse dynamics for robotic manipulation: given a predicted future visual state, infer the motor command to reach it via cross-attention retrieval over a demonstration memory bank.
  • Achieved 6.5x reduction in action prediction error vs. direct neural network baseline on LIBERO-Spatial with only 25 demonstrations (val MSE 0.131 vs. 0.852).
  • Extended to RL fine-tuning in simulation, reaching 25% task success without retraining from scratch.
LEO Ground-Station Network Design Spring 2026
INDENG 230 – Supply Chain Analytics
  • Modeled LEO satellite ground-station placement as a supply chain network design problem using MILP to trace cost-coverage Pareto frontiers across infrastructure constraints including backhaul availability and geopolitical feasibility.
  • Population-weighted coverage reached 43.3% vs. 20.8% uniform at identical geometry; geopolitical constraints reduced median distance to actual Starlink gateways from 7,154 km to 881 km.
Understanding Customer Churn for BizGrow Fall 2025
Google Hackathon
  • Led team analyzing SaaS customer journey data with Python/Pandas; applied keyword clustering and segmentation to isolate a high-revenue cohort responsible for ~80% of churn revenue loss.
  • Validated findings with Causal ML models (Double ML, Uplift); delivered executive presentation addressing $29M in at-risk contracts.

Volunteerism

California Academy of Sciences – San Francisco, CA Dec 2022 – May 2023
Scientific Computing Volunteer
  • Built a Flask/Python dashboard with Bash collection agents to monitor CPU/GPU usage, job queues, and storage across research clusters; informed equitable compute scheduling for the Director of Scientific Computing.

Skills & Interests

LanguagesPython, SQL, TypeScript, JavaScript, C++, HTML/CSS
Libraries & FrameworksPandas, NumPy, Scikit-learn, PyTorch, LangChain, FAISS, Sentence-Transformers, Spark
Platforms & ToolsClaude Code, Google Cloud SQL, Databricks, Firebase, MongoDB, Git, Dash/Plotly
MethodsRAG, Embedding-Based Search, LLM Agents, LLM Fine-Tuning, MLOps, Model Deployment & Versioning, Prompt Engineering, Multi-step Reasoning, Hallucination Control, Vector Search, Statistical Modeling, ETL, REST APIs, A/B Testing
OtherInstructing & Public Speaking, Cross-functional Team Leadership, GIS, UX/UI Design
InterestsScuba Diving, Motorcycling, Photography, Aquarium Design