James Murray
Senior technical builder, systems architect, and software developer with more than 40 years of programming and product-building experience. I design and ship complete systems, from data collection and preparation to vector databases, RAG search, APIs, production deployment, and public-facing PHP websites.
My work combines deep software experience, recovery-industry knowledge, and a practical ability to turn complex ideas into working systems that help real people.
Builder of complete systems
I do not work only at the idea level. I build working systems across the full stack: research, data extraction, cleaning, JSON and CSV preparation, vector ingestion, search logic, API design, PHP integration, deployment, testing, logging, backup strategy, and production rollout.
AI knowledge systems and vector search
I create private, grounded AI search systems that rely on curated source data rather than loose public-web chat output. That includes semantic search, therapist and rehab discovery tools, RAG pipelines, and secure production safeguards.
Recovery, health, and mission-driven platforms
My most visible work has focused on addiction recovery, therapist discovery, information clarity, and turning large bodies of content into usable guidance for people in difficult moments.
Core competencies
The value I bring is not one isolated tool. It is the ability to combine strategy, programming, infrastructure, and production execution into a coherent technical system.
AI, vector search, and data systems
- Private AI search systems and source-grounded answer generation.
- Vector databases including Weaviate, Qdrant, Chroma, Milvus, and Pinecone.
- Embeddings, chunking strategy, metadata structuring, and retrieval tuning.
- RAG APIs for therapist, rehab, and knowledge-base workflows.
- Structured testing, regression validation, and search-quality evaluation.
Application and web engineering
- PHP, Python, JavaScript, SQL, Bash, PowerShell, and Linux command-line workflows.
- FastAPI, Uvicorn, API proxies, secure headers, rate limiting, and logging.
- Custom PHP pages, dynamic search interfaces, protected image delivery, and form handling.
- JSON, CSV, JSONL, SQL export/import workflows, and content pipelines.
- Mobile-aware front-end integration, UX organization, and information architecture.
Infrastructure and deployment
- Ubuntu VPS setup, service configuration, systemd management, and remote administration.
- SSH-based deployment, GitHub-based version control, staging and rollback planning.
- Monitoring, health checks, backups, restore bundles, and production candidate packaging.
- Hosting-server integration for public web pages, API proxies, and protected assets.
- Secure production rollout with controlled soft-launch procedures.
Business and solution delivery
- Technical translation for founders, operators, and non-technical stakeholders.
- Programmatic SEO systems and scalable city / province content frameworks.
- Recovery-focused information products and professional directory strategy.
- Consulting, contract builds, production rescue, and system modernization.
- End-to-end ownership from concept through deployment and stabilization.
Selected technical experience
Founder / Architect — AddictionTube knowledge systems
Designed and built a private therapist and rehab search architecture that connects research, data pipelines, vector databases, FastAPI services, secure PHP proxies, and public-facing search pages. The system supports search by concern, location, professional profile, and direct therapist questions.
- Built curated knowledge pipelines from real provider websites and structured profile information.
- Created staged and production API paths, pagination logic, city directory browsing, and secure image token handling.
- Implemented operational safeguards including regression testing, rate limiting, health checks, backups, and restore bundles.
- Developed standalone city and province therapist pages designed for discoverability and public trust.
AI and vector database consultant
Worked on applied AI search and knowledge projects where the goal was not generic chatbot behavior, but practical, source-grounded answers. This included design of vector schemas, retrieval flows, ingestion logic, and content structuring for domain-specific search systems.
- Mapped business questions into vector-search architectures and usable interfaces.
- Built and tested ingestion-ready datasets using CSV, JSON, JSONL, SQL, and scraped source material.
- Created structured metadata and chunking logic to improve answer quality and control hallucination risk.
Software developer and systems builder
Long-term experience across custom software, data systems, websites, automation, and problem-solving for real-world operations. My work spans early software development, custom web systems, data integration, analytics tools, and modern AI-enabled information systems.
- Designed and built software products, websites, utilities, and structured information systems across decades of technology change.
- Adapted continuously from traditional software and web engineering into modern AI, vector database, and production API environments.
- Comfortable working independently across architecture, coding, deployment, and troubleshooting.
What I built for the AddictionTube therapist search system
This work is a strong example of the level I can operate at: not just coding a page, but creating the full operational system behind it.
Data creation
- Local PC research and extraction workflows.
- Python scripts for crawling, preparation, cleaning, and transformation.
- Structured provider records, JSONL chunk files, and import-ready artifacts.
VPS architecture
- Ubuntu VPS with FastAPI / Uvicorn application services.
- Weaviate vector database collections and controlled API endpoints.
- Structured logging, health monitoring, and backup / rollback workflow.
Public integration
- Hosting-server PHP search page and API proxy.
- Protected therapist image delivery and tokenized access.
- Public search, city browsing, pagination, and ask-a-therapist-question flow.
Tools, platforms, and technologies
Languages
PHP, Python, JavaScript, SQL, Bash, PowerShell.
AI and vector stack
OpenAI APIs, embeddings, RAG workflows, Weaviate, Qdrant, Chroma, Milvus, Pinecone.
Web and APIs
FastAPI, Uvicorn, REST endpoints, PHP proxies, JSON APIs, secure headers, rate limiting.
Infrastructure
Ubuntu VPS, systemd services, SSH, remote access, logging, monitoring, backups.
Data formats
CSV, JSON, JSONL, SQL dumps, structured exports, metadata-enriched content chunks.
Workflow tools
GitHub repositories, staging pipelines, regression test scripts, deployment runbooks.
Availability
Consulting
Architecture reviews, AI strategy, private search systems, data-to-knowledge workflows, and search-quality improvement.
Contract development
Hands-on implementation in PHP, Python, FastAPI, vector databases, structured content systems, and production integration.
Production rescue and rebuilds
Fixing broken systems, stabilizing search stacks, rebuilding pipelines, and bringing projects from concept to working deployment.
For opportunities, consulting, or collaboration, use the contact page or review the broader technical portfolio for project evidence and scope.