Kieran Gookey

• Re-engineered production chatbot system with advanced RAG pipeline using pgvector and context-aware query transformations (incorporating error messages, user selections). • Implemented sophisticated multi-format document chunking (HTML, PDFs, transcripts) and self-hosted CPU-based reranker on Kubernetes for cost-effective inference. • Built custom chat history management with automatic summarization and metadata handling, full observability through Langfuse. • Developed financial agent system using Pydantic AI and LangGraph to interface with Jedox’s proprietary cube, analyzing sensitive financial data to generate business unit performance insights. • Consulted on cost-effective microservice AI architecture best practices. • Skills: Python, RAG, pgvector, Langfuse, Llama Server, Pydantic AI, LangGraph, Kubernetes, Document Processing, Agentic AI, Financial Data, Cost Optimization

• Architected and deployed a sophisticated document intelligence solution using multi-modal AI models to extract structured and unstructured data from complex documents with high accuracy. • Developed custom methods to accurately identify tables within images and read diverse content, including handwritten data. • Utilized both traditional computer vision techniques and multi-modal AI models to enhance the accuracy and reliability of data extraction. • Engineered a cost-effective model serving architecture on shared GPU resources, ensuring scalable and performant inference. • Skills: Python, Computer Vision (OpenCV), Multi-modal Models, Document Analysis, Data Extraction, Model Deployment & Scaling, GPU Resource Management

• Replaced a manual, non-scalable 8-VPS scraping process with a modern, event-driven GKE architecture. • Built and deployed new infrastructure using Terraform and established automated GitOps workflows with ArgoCD. • Implemented KEDA for efficient, event-based scaling of scraping workloads. • Optimized resource utilization and cost-effectiveness by configuring efficiently sized Chromium pods/nodes and leveraging Spot Instances with automated pod rollover. • Skills: Web Scraping, Terraform, Kubernetes (K8s/GKE), GCP, GitOps (ArgoCD), KEDA, Infrastructure Architecture, Cost Optimization, Browser Automation, Supabase

• Led the greenfield architecture of a high-volume data acquisition platform, processing over 3 million e-commerce pages daily. • Engineered and scaled production-grade RAG pipelines using Llama3 and Weaviate for dynamic data extraction, serving millions of daily requests. • Designed and deployed a predictive XGBoost model for intelligent proxy orchestration, increasing data acquisition efficiency by 500%. • Built and maintained complex Spark pipelines on AWS to process over 10 million items daily. • Built efficient data orchestration pipelines in Go. • Architected and led multiple other large projects as part of a fully remote team. • Principal software engineer that took over after a failed agency and built a team of 5 engineers through multiple funding rounds. • Skills: Python, Go, TypeScript, AWS, GCP, RAG (Llama3, LangChain), XGBoost, Vector Databases (Weaviate), Kubernetes (EKS), Spark, Data Pipelines, MLOps

• Set up a full development workflow and production-ready environment in AWS for the largest building society in the UK. • Led multiple teams in the development and deployment of their first cloud-based projects. • Completed a full migration from on-prem to AWS. • Successfully designed and implemented cloud-native development workflows. • Performed above expectations leading to a contract extension from 6 to 12 months.

• Designed and built scalable containerised backend microservices for complex mortgage products and calculations. • Integrated with as well as completed contracts for the largest UK lenders/brokers. • Led the implementation of new testing tooling.

• Built bespoke software solutions for worldwide clients, including EMEA, APAC, and AMER regions. • Developed visualizations for enormous amounts of financial data for a branch of the New Zealand government. • Created dynamic and customisable surveys for the largest UK further education institutions. • Built flexible data import and aggregation services used in a variety of contracts and sales pitches. • Skills: C# (.NET), MSSQL, JQuery, Azure, Tableau, Data Visualisation