Abishek Murali

About Abishek Murali

I work at Canvass, where I lead the design and development of the firm’s ML and the GenAI agentic platform. Recently, I played a foundational role in architecting MONET — Canvass’s flagship AI agent framework — through direct collaboration with Microsoft\'s OpenAI lab. My work involved bridging advanced GenAI models and real-world operational use cases, enabling enterprise users to interact with AI agents through intuitive interfaces and backend orchestration. I led the buildout of reusable agent architectures, deployed reinforcement learning systems for critical industrial optimization, and pioneered NLP-driven interfaces like natural language to SQL translation.I\'ve had experience in building products with machine learning at it\'s core to provide value to Canvass\' customers. I\'ve had extensive experience in collaborating with customers, product, devops, sales and marketing to build and ship out these products.My career began in business intelligence and NLP which allowed me to establish solid principles in being able to translate customer requirements into solutions that can work at scale.

Experience

1. Lead Machine Learning Engineer

   Canvass AI

   Aug 2018 — Present · Toronto, ON, CA

   Led ML + GenAI projects for industrial time-series/operations; collaborated with product and customers to design from ground-up- Developed reusable LLM agent templates, safety guardrails, and orchestration flows to accelerate enterprise adoption; used langchain and llama-index for development and langsmith for observability- Reduced ML time-to-prod from 1 month to 1 week via MLOps pipelines on AKS, Kubeflow, and CircleCI- Architected and built MONET, Canvass AI’s GenAI agent platform; onboarding months → days; collaborated with Microsoft\'s OpenAI lab- Created natural language to SQL RAG interfaces (used Qdrant vector store) enabling self-serve analytics across departments- Built P & ID parsing and reconciliation workflows reducing turnaround time from weeks to hours- Delivered co-gen optimization saving $2M and launched an RL control agent for a blast furnace (staged rollout with safety guardrails)- Slashed inference latency 6x and mentored junior engineers on infrastructure, model design, and production delivery- Helped team integrate Datadog to get visibility into applications running on production

Education

• B. M. S. College of Engineering

  Bachelor’s Degree, Electrical, Electronics and Communications Engineering

  2011 — 2015

• The University of British Columbia

  Master's degree, Data Science

  2017 — 2018

Skills

• Programming
• Data Visualization
• C++
• Python
• .net
• Microsoft Bi Suite
• Tableau
• Sql
• Microsoft Office
• Matlab
• Javascript
• Microsoft Excel
• Microsoft Sql Server