Mohammed Farhan Baluch
AI Engineer, Toronto, Canada
farhanbaluch1301@gmail.comlinkedin.com/in/farhan1301github.com/farhan1301farhanbaluch.com
AI Engineer at ML Arteka (formerly mobileLIVE), an AI and digital consultancy. I design multi-agent LLM systems for enterprise clients in telecom and banking, then build the evaluation, test automation and rollout that make them trustworthy. Before that, an MSc in AI with five peer-reviewed publications.
Experience
AI Engineer, ML Arteka (formerly mobileLIVE)
Oct 2024 – present- Leading enterprise adoption of an agentic test-architecture framework: risk-based test design, AI-generated tests, traceability and CI/CD quality gates.
- Designed an evaluation pipeline that scores agents on 10+ quality dimensions with contract-based scenarios and OpenTelemetry traces.
- Built an orchestrator-worker agent framework that autonomously tests a production voice assistant, plus a design agent that writes branch-covering test cases in under 10 minutes per flow.
- Migrated a Visual Basic and Crystal Reports estate to Java 21 with spec-driven, multimodal agents; 70% faster than planned.
- Built an LLM crawler across 20+ bank websites at 99.3% accuracy, turning two weeks of market analysis into under an hour.
- Led a five-agent estimation workflow that cut manual estimation effort by 80%.
- Drew Azure and AWS AI solution architectures for pre-sales; helped win 3 enterprise clients across 10+ proofs of concept.
Graduate Research Assistant, University of Windsor
Sep 2022 – Sep 2024- Built a graph-augmented TD3 agent (Node2Vec embeddings) that outperformed RL baselines on Dow 30 data.
- Improved risk-adjusted returns with sector clustering and technical indicators, and made the agent's strategy auditable with post-hoc visualization.
Technical Assistant, Vector Institute
Jan 2024 – Apr 2024Helped small and mid-sized businesses take AI use cases from problem definition to prototype in three months.
Graduate Teaching Assistant, University of Windsor
Sep 2022 – Apr 2024Tutorials and labs in computer architecture and systems programming for 100+ students.
AI Developer, Agriculture and Agri-Food Canada
May 2023 – Aug 2023Computer vision for early crop-disease detection, and a policy chatbot for farmers.
Skills
- Agents and evaluation
- Multi-agent orchestration, Agent evaluation and LLM-as-judge, Copilot custom agents and skills, Claude Agent SDK, LangGraph, Model Context Protocol, GraphRAG and RAG, Realtime voice agents
- Quality engineering
- Playwright, Risk-based test strategy, Coverage and quality metrics, Conversational AI testing
- Platform and delivery
- Agent packaging and rollout, Spec-driven development, Legacy modernization, Developer experience tooling, AI governance and standards
- Languages and cloud
- TypeScript and Node.js, Python, Java and Spring Boot, Azure (AKS, ML, OpenAI, Databricks), AWS (Bedrock, SageMaker, Lambda)
Education
MSc Computer Science, AI specialization, University of Windsor
2022 – 20244.0 GPA, Ontario Graduate Scholarship. Thesis on graph-augmented reinforcement learning for stock trading.
BTech Computer Science and Engineering, Vellore Institute of Technology
2018 – 2022Also
- Small-model fine-tuning. Fine-tuned Llama 3.2 with QLoRA on consumer hardware for structured data extraction. Exact-match accuracy went from about 25% to about 75%.
- RAG evaluation lab. An open workspace for benchmarking RAG pipelines before production. Cut manual evaluation time by 60%.
- Research. 5 peer-reviewed papers in fraud detection, computer vision and healthcare prediction.
- Certifications. AI for Medicine specialization; Introduction to Generative AI; Reinforcement learning in Python.
- Languages. English (professional), Hindi (native).