> _
Data Scientist & ML/GenAI Engineer with 3 years building 20+ professional dashboards, ETL pipelines, and production AI systems — multi-agent pipelines, RAG architectures, and quantitative trading strategies. Research Assistant at ASCI Lab × Siemens, Trent University.
Data Scientist & ML/GenAI Engineer with 3 years building 20+ professional dashboards, ETL pipelines, and production AI systems — multi-agent pipelines, RAG architectures, and quantitative trading strategies.
My work sits at the intersection of quantitative finance and industrial AI. Currently a Research Assistant at Trent's ASCI Lab in collaboration with Siemens, developing multi-agent systems for building automation and energy optimization.
I care deeply about systems that don't just predict — they reason over context, adapt to new information, and take meaningful actions.
"Building AI systems that don't just predict —
they reason, adapt, and act."
1class Dylan: 3 role = [ 4 "Data Scientist", 5 "ML/AI/GenAI Engineer", 6 "Agentic AI Builder", 7 ] 9 location = "Toronto, ON 🇨🇦" 11 education = { 12 "MSc": "Big Data Analytics", 13 "gpa": "95% · Highest Honor", 14 "BEng": "Financial Technology", 15 } 17 currently = "RA @ ASCI Lab × Siemens" 19 focus = [ 20 "Multi-Agent AI", 21 "RAG Pipelines", 22 "Production ML", 23 "Quant Trading", 24 ] 26 open_to = "Contract & Full-Time · Data Sci · GenAI · Analytics" 28 def build(self): 29 # systems that reason, adapt, and act 30 return "AI × Finance → Impact"
Production systems, research projects, and open-source tools.
Production-grade AI for occupancy analysis & energy optimization
Flask platform with a Chart.js admin dashboard, real-time SSE pipeline streaming, and a human-in-the-loop confirmation loop. Multi-LLM routing across Claude, GPT-4o, and Ollama/DeepSeek. Self-healing execution agents backed by ChromaDB + FAISS RAG.
System Architecture
Multi-AI Autonomous Trading Agent — Stocks & Forex
End-to-end autonomous trading system combining TCN-LSTM hybrid signal generation with multi-agent orchestration. Live execution via Oanda API (Forex) and IB Gateway (Stocks). Walk-forward validation, vol-adjusted stop-loss, and drawdown controls.
Multi-Agent Architecture
Spark + Kafka + LLM Explanations
Scalable scam detection pipeline for phone dialogue transcripts. Spark ML for classification, Kafka for streaming ingestion, LLM-generated natural language explanations per flagged call.
Azure × Databricks · CNN-LSTM-Transformer
Dual-model forecasting for Toronto traffic collisions under weather conditions. CNN-LSTM-Transformer architecture with SHAP explainability and Power BI dashboard. Deployed on Azure Databricks.
Power BI Dashboard
React + Flask · LangChain · End-to-End
Full-stack RAG system with a React frontend and Flask backend. LangChain orchestration for document ingestion, chunking, embedding, retrieval, and generation.
DQN + Classical Search · Full-Stack Flask
Web-playable Mancala game with five AI opponents: DQN, Minimax, Alpha-Beta pruning, MCTS, and an advanced heuristic agent.
ML Factor Models · Canadian Equity Market
ML-driven factor models combined with fundamental signals for the Canadian equity market. Quantamental approach blending systematic signals with macro context.
Django · Crypto News Platform
Full-stack crypto news and community platform built with Django. User authentication, dynamic content rendering, and responsive design.
Open to contract & full-time roles in Data Science, Data Engineering, ML/AI/GenAI Engineering, and Analytics Engineering. Toronto & Remote.