Spencer G.Private profile: Upwork shows its earnings only to signed-in members, so we check it every few weeks, not daily. MRR here is a rough estimate (Upwork's six-month earnings ÷ months active). Last read Jul 17, 2026.

𝗦𝗲𝗻𝗶𝗼𝗿 𝗔𝗜 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿 → 𝗗𝗮𝘁𝗮 𝗦𝗰𝗶𝗲𝗻𝗰𝗲 → 𝗙𝘂𝗹𝗹 𝗦𝘁𝗮𝗰𝗸 → 𝗶𝗢𝗦 6 years building things that work. Not demos. Production systems. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ ⚡ 𝗪𝗛𝗔𝗧 𝗜 𝗔𝗖𝗧𝗨𝗔𝗟𝗟𝗬 𝗗𝗢 I turn messy problems into deployed solutions. As a Founding AI Engineer, I've owned entire product lifecycles—from "we need AI for this" to "it's live and processing millions of requests." Not a prompt-wrapper developer. I build the infrastructure underneath. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🧠 𝗔𝗜 / 𝗟𝗟𝗠 / 𝗔𝗚𝗘𝗡𝗧𝗦 The stuff everyone's talking about—I've shipped it: LangChain + LangGraph → Multi-agent systems that actually reason RAG Pipelines → Your data + LLMs = answers that don't hallucinate Voice AI → Real-time conversational agents, not chatbot toys Vector Search → Pinecone, FAISS, Chroma, Weaviate—pick your weapon Models I work with daily: OpenAI • Claude • Gemini • Llama • Mistral The unsexy but critical parts: ↳ Prompt engineering that survives edge cases ↳ Embedding strategies that scale to millions of documents ↳ Evaluation frameworks (F1, precision, recall—metrics that matter) ↳ Cost optimization and guardrails for production deployment ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🔬 𝗗𝗘𝗘𝗣 𝗟𝗘𝗔𝗥𝗡𝗜𝗡𝗚 / 𝗖𝗟𝗔𝗦𝗦𝗜𝗖𝗔𝗟 𝗠𝗟 Before LLMs were cool, I was building: → Siamese Networks (Sentence-BERT + CLIP/ResNet50 for product matching) → Triplet loss optimization + semi-supervised learning loops → Production pipelines: TensorFlow, PyTorch, Keras → Real-time inference at scale with FAISS indexing I understand why models work, not just how to call them. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 📊 𝗗𝗔𝗧𝗔 𝗦𝗖𝗜𝗘𝗡𝗖𝗘 & 𝗔𝗡𝗔𝗟𝗬𝗧𝗜𝗖𝗦 Healthcare AI (HIPAA/GDPR compliant): → 50+ biometric signals → predictive models for early disease detection → UTIs, kidney stones, diabetes, heart failure, preeclampsia → Real-time anomaly detection + automated alerting systems The boring stuff that makes projects succeed: ↳ Feature engineering at scale ↳ Statistical rigor (hypothesis testing, A/B experiments) ↳ Data pipelines and ETL workflows that handle messy inputs ↳ Visualization and reporting stakeholders actually understand ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🔧 𝗙𝗨𝗟𝗟 𝗦𝗧𝗔𝗖𝗞 𝗘𝗡𝗚𝗜𝗡𝗘𝗘𝗥𝗜𝗡𝗚 𝗕𝗮𝗰𝗸𝗲𝗻𝗱 Node.js / Express → FastAPI / Flask → REST APIs that don't break PostgreSQL / MongoDB / Redis → Data that's queryable and fast AWS / GCP / Heroku → Deployed, monitored, scaled 𝗙𝗿𝗼𝗻𝘁𝗲𝗻𝗱 React.js → D3.js → Dashboards and data viz that tell stories 𝗗𝗮𝘁𝗮 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 ETL pipelines → Real-time streaming → Low-latency processing 𝗪𝗲𝗯𝟯 / 𝗕𝗹𝗼𝗰𝗸𝗰𝗵𝗮𝗶𝗻 Blockchain integration (EVM/UTXO) → Salesforce CRM sync Wallet connections → Real-time transaction mapping ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 📱 𝗶𝗢𝗦 𝗗𝗘𝗩𝗘𝗟𝗢𝗣𝗠𝗘𝗡𝗧 Apps in the App Store. Not tutorials. SwiftUI / UIKit → Production-grade code that passes review Apple Watch → Companion apps that extend the experience Core ML → On-device intelligence without server roundtrips Xcode Cloud → CI/CD pipelines that ship reliably ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🛠 𝗧𝗢𝗢𝗟𝗦 𝗜 𝗥𝗘𝗔𝗖𝗛 𝗙𝗢𝗥 𝗔𝗜/𝗠𝗟 Python • TensorFlow • PyTorch • HuggingFace • Scikit-learn 𝗟𝗟𝗠 LangChain • LangGraph • OpenAI • Claude • RAG • Voice AI 𝗩𝗲𝗰𝘁𝗼𝗿 𝗗𝗕 Pinecone • FAISS • Chroma • Weaviate • Qdrant 𝗕𝗮𝗰𝗸𝗲𝗻𝗱 Node.js • FastAPI • Flask • Express • REST APIs 𝗗𝗮𝘁𝗮 PostgreSQL • MongoDB • Redis • Pandas • NumPy 𝗙𝗿𝗼𝗻𝘁𝗲𝗻𝗱 React.js • D3.js • JavaScript 𝗶𝗢𝗦 Swift • SwiftUI • Core ML • Xcode • Apple Watch 𝗖𝗹𝗼𝘂𝗱 AWS • GCP • Docker • CI/CD ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ ✦ 𝗛𝗢𝗪 𝗜 𝗪𝗢𝗥𝗞 → I own problems end-to-end → I ship incrementally (progress, not excuses) → I communicate like an engineer (clear, direct, no fluff) → I optimize for outcomes, not hours billed ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ If you need AI that works in production—not just notebooks—let's talk. If you need full-stack with real engineering behind it—let's talk. If you need an iOS app that ships and scales—let's talk.