Patrick Prunty

Patrick Prunty

Senior Web Developer in Dublin, Ireland, he/him

Work Experience

2020 — Now
Dublin, Ireland

Worked across three organizations, leading full-stack development and optimizing performance for large-scale web applications, with additional experience in AI/ML engineering, focusing on scalable model deployment, cost optimization, and reproducible machine learning workflows. Key contributions:

  • Led development of two fully authenticated web applications using OAuth2.0, SSO, and passwordless (SMTP) authentication. Secured APIs with JWT-based access/refresh tokens and RBAC-controlled admin access.

  • Optimized a data-intensive interface handling 50k+ rows and 15+ filters in Next.js, leveraging caching (Redis/React), skeleton loaders, and virtualized scrolling. Applied similar optimizations to Recharts/Plotly, cutting dashboard load times by 40% and boosting Lighthouse scores from 55% to 90%+.

  • Built 12+ modular React components for an internal shared library, featuring modern animations (e.g., animated modals, skeleton loaders, OTP inputs, tooltips, bottom navigation bar for mobile devices, lucide-react icons support). Documented in Storybook and tested with RTL/Jest. Library usage rose by 25% after a company-wide announcement of the new major release.

  • Contributed to an internal Python SDK for AI/ML models in Databricks, improving execution time by 25% and cutting compute costs by 40%. Enhanced reproducibility with Databricks Data Catalog and standardized runs using cluster policies (Photon, GPU, auto-scaling). Adopted by 6+ data science teams, running 40+ disease prediction models reliably on a daily/monthly basis.

  • Developed cloud-based microservices on Azure (AKS, Functions, Blob Storage, Event Hub, Key Vault, Container Registry) and AWS (ECS, ECR, S3, API Gateway, Route53, VPC), ensuring event-driven communication via Kafka. Built RESTful APIs with Node.js (TypeScript) and FastAPI (Python) for large-scale data ingestion and SQL storage via ORMs. Collaborated with DevOps teams to automate infrastructure with Terraform, Jenkins, and GitHub Actions for CI/CD.

  • Partnered with stakeholders and product managers to align development with business goals and user feedback on a 2-week sprint cadence. Led bi-weekly demonstrations of new features, refined roadmaps, and prioritized user stories to accelerate high-impact releases.

Languages: SQL, Python*, Java, TypeScript*, JavaScript/HTML/CSS*, Bash*.

Frameworks/Libraries: FastAPI*, Django, Spring, Node, React*, Next.js*, TailwindCSS, PySpark.

Tools: Docker, GitHub Actions CI/CD*, Terraform*, Redis, Databricks*, Azure*, AWS.

Note: Skills marked with an asterisk (*) denote advanced industry experience.

Education

2019 — 2020
Dublin
  • Developed expertise in C++17/20, adhering to STL and community standards.
  • Built advanced computational models, including an infectious disease simulator (survival models), a financial options contract pricing engine, and a reinforcement learning model for playing chess.
  • Ran large-scale computations using MPI, OpenMP, and NVIDIA CUDA, deploying distributed processing on Trinity’s supercomputer.
2016 — 2019
Dublin, Ireland
  • Specialized in partial differential equations, stochastic models, and advanced linear algebra.

  • Gained programming experience through introductory Python and C++ modules.

  • Conducted research on modernist writers such as James Joyce and T.S. Eliot using post-structuralist analysis.

  • Achieved CEFR B1 in Portuguese, enhanced by a summer program at NOVA Lisbon; basic proficiency in Spanish.

Side Projects

2025

Educational channel exploring computational mathematics and computer science through storytelling and visual animations (videos coming soon).

2025

Providing insight into trending films and television series through video essays (videos coming soon).

2025

High-performance blog built with Next.js 15 and MDX, where I share insights on life, technology, and personal projects.

Projects

2025

An open-source Next.js template featuring progressively styled, reusable React components and layouts. Includes built-in authentication, SEO optimization, and a clean foundation for SaaS products. Ideal for developers seeking a minimalist yet extensible starting point (work in progress).

2025

Interactive Manim environment for seamless mathematical animations, replicating YouTube channel 3Blue1Brown’s workflow through code highlighting and keyboard binding shortcuts.

2025

A foundational C++ project providing a connection to Coinbase’s L3 market data feed, storing multi-market data in an efficient orderbook and offering performance benchmarks (work in progress).

2025

An autonomous AI agent capable of piloting the creation of an entire coding given a project outline using DeepSeek open-source LLMs. The user simply sits back and determines whether to accept/reject iterations (work in progress).

Awards

2019
James Joyce Scholarship from University College Dublin

Awarded a scholarship for the James Joyce Summer School at Boston College, recognizing outstanding critical and analytical writing in English Literature final year results.

LinkedIn
Strava