class FarhanRahman:
role = "Machine Learning Engineer @ Vanguard"
location = "Toronto, Canada 🇨🇦"
education = [
"MEng Computer Engineering — University of Toronto (GPA 3.9 / 4.0)",
"B.Tech ECE — VIT Chennai (GPA 3.89 / 4.0)",
]
expertise = ["MLOps & LLMOps", "Agentic AI & RAG Systems", "Fraud Detection", "Scalable GenAI Infrastructure"]
companies = ["Vanguard", "Canadian Tire Corporation", "Rewind Software", "Nightingale AI", "Cilia AI"]
currently = "Shipping production-grade ML/LLM infrastructure that saves $500K+/year"ML Engineer specializing in MLOps, LLMOps, and enterprise GenAI platforms — from architecting RAG pipelines and agentic workflows to shipping fraud detection systems at scale.
| Project | Stack | Impact |
|---|---|---|
| 🏦 New Account Fraud Detection @ Vanguard | AWS Lambda · Step Functions · ECS Fargate · EventBridge | 💰 $500K+/yr saved |
| 🤖 LLM Risk Commentary Engine | LangChain · AWS Bedrock · Azure OpenAI · HITL feedback | 🎯 Production-grade |
| 🔍 ML Model Observability Framework | Arize AI · Prometheus · Grafana | 📊 Fleet-wide coverage |
| 📄 Multimodal Document Intelligence (MIRANDA) | LangChain · Vision LLMs · vg-document-agent | 🔬 Image & table analysis |
🏦 Vanguard — Machine Learning Engineer, Specialist | Aug 2025 – Present
- 🔒 Deployed a New Account Fraud Detection system using AWS Lambda, Step Functions, EventBridge & CI/CD — saving $500K+/year
- 🤖 Built a modular LLM risk commentary system using LangChain, Bedrock, SageMaker & Azure OpenAI with LLM-as-Judge and Human-in-the-Loop feedback
- ⚡ Developed a real-time fraud detection FastAPI inference engine on ECS Fargate with RBAC authentication
- 📄 Built a reusable Python document intelligence library for multimodal LLM analysis (MIRANDA / vg-document-agent)
- 📊 Standardized ML model onboarding & observability framework via Arize AI across the production model fleet
🛞 Canadian Tire Corporation — AI ModelOps Engineer | Nov 2024 – Aug 2025
- 📈 Scaled enterprise AI/ML platform to 10,000+ models using Apache Spark and cost-optimized spot instances
- 🔗 Built LLMOps framework with LangSmith, Azure Databricks, MLflow, and full CI/CD via Azure DevOps
- 📡 Deployed Prometheus + Grafana monitoring for real-time model health and drift detection
- 🕸️ Designed end-to-end RAG systems and Agentic AI architecture using LangGraph, LangChain, Azure AI Foundry & MCP servers
- 🌐 Drove organization-wide GenAI adoption across sales and supply chain domains
⏪ Rewind Software — Full Stack AI Engineer | Sep 2023 – Nov 2024
- 🧑💻 Built a custom coding assistant with fine-tuned GPT-4, PaLM & Gemini Pro using CoT, SFT & RLHF
- 🚀 Reduced time-to-production by 25% with proprietary RAG and automated backup restore services
- 🔴 Built automated LLM red-teaming pipelines for jailbreaks, prompt injection, data leakage & reproducibility
- 🔐 Deployed secure LLMOps infrastructure on AWS (API Gateway, IAM, Bedrock, SageMaker, Lambda, S3)
🏥 Nightingale AI — Machine Learning Engineer | May 2022 – Sep 2023
- 🧍 Built an end-to-end pose estimation platform using MediaPipe Holistic for clinical ROM analysis
- ⚡ Deployed real-time ML inference for live video on AWS ECS Fargate & EKS (high-throughput, low-latency)
- 🔔 Improved incident response time by 40% with cloud ML pipelines on EC2, S3, Lambda & CloudWatch
🎨 Cilia AI — Machine Learning Engineer | Nov 2021 – Jun 2022
- 🖼️ Designed end-to-end ML architecture for artwork detection using Azure DevOps, Azure ML Studio & Custom Vision
- 🔥 Optimized large-scale data pipelines with Apache Spark & Ray for distributed model training
- 📦 Improved deployment efficiency by 30% via Kubeflow, Docker, Terraform & MLflow automation
| Badge | Certification | Issuer | Valid |
|---|---|---|---|
| ☁️ | AWS Certified Solutions Architect – Associate | Amazon Web Services | Feb 2026 – Feb 2029 |
| 🤖 | Azure AI Engineer Associate | Microsoft | Aug 2025 – Aug 2026 |
| 🔒 | Security Champion – Level 1 | Vanguard | Feb 2026 – Present |
- 📌 AI Agents That Edit Your Blogs and Ping Your Team: LangGraph + Slack in Action
- 📌 Configuring your CUDA drivers for custom object detection (Linux)
- 📌 Model Order Selection for Neural Data
- 📌 Data Beyond Perception
- 📌 How Ontario's Mobility is Shaping Up Since Reopening
"The best models are the ones that actually ship."
💬 Open to collaborations in MLOps, LLMOps, Agentic AI, and production GenAI systems. Feel free to reach out via LinkedIn or Email.