About

I lead engineering teams that ship enterprise AI in production — from multi-agent platforms to edge inference — across cloud, on-prem, and air-gapped environments.

Engineering Manager at Iterate.ai, where I lead a 12-engineer cross-functional organization building Generate, the company's flagship enterprise AI platform. In under two years I grew from data scientist to running the org behind a product recognized as AI Product of the Year by both the Pinnacle and TMC awards, with partnerships spanning IBM, Intel, AMD, and HP.

I pair that leadership with deep hands-on AI/ML: I hold an MASc in Electrical & Computer Engineering from McMaster University (thesis on SciRAG, a retrieval-focused fine-tuning strategy for scientific documents), 4 filed patents, and peer-reviewed publications with 245 citations.

RAG · Multi-Agent Systems · Edge & On-Prem Inference · Document Intelligence · Team Leadership

Impact at a glance

12+
Engineers led
(scaled from 6)
2×
AI Product of the Year, 2025
4
Patents filed
(US)
245
Research citations

Experience Full history on LinkedIn →

Engineering Leadership

Engineering Manager

May 2025 — Present

Iterate.ai — Toronto, ON

  • Scaled the org from 6 to 12 engineers across frontend, backend, DevOps, and AI/ML — led 6 hires, promoted 2 ICs to Senior Applied AI Engineer, and mentored 4 engineers plus 3 interns.
  • Own engineering for Generate end-to-end across cloud, on-prem, and edge (AWS, IBM Cloud, Intel/AMD/NVIDIA) — enabling both SaaS and air-gapped enterprise deployments.
  • Launched Generate for healthcare revenue recovery via an IBM reseller partnership with Hutchinson Regional Healthcare, featured in IBM's product blog.

Engineering Manager (Contract)

Sep 2024 — May 2025

Iterate.ai — Remote, Canada

  • Took ownership of a 5-engineer team and defined the roadmap that made Generate Iterate.ai's primary enterprise revenue product.
  • Architected an event-driven microservices platform (Kafka, FastAPI, LangGraph) for high-throughput AI workloads; established the engineering operating model later inherited by the full-time team.
Applied AI & Research

Data Scientist

Nov 2023 — Sep 2024

Iterate.ai — Hamilton, ON

  • Led a 5-engineer team building edge-deployed AI for private document search on Intel AI PCs, shipping through the Intel Software Advantage Program.
  • Architected an Outlook RAG plugin for email automation — presented at the Intel Vision 2024 keynote.
  • Built a novel table-extraction system (Non-Maximum Suppression + Vision Language Models); filed US Patent App. 18/590,347.

Junior ML Engineer → ML Research Engineer

2021 — 2023

ExentAI & Simplyfai — Sri Lanka (Remote)

  • ExentAI: built document-digitization and multilingual OCR pipelines (Tamil/Sinhala/English) on AWS with Docker & Kubernetes; shipped LLM pipelines for sentiment, emotion, and hate-speech detection.
  • Simplyfai: built fault-classification and anomaly-detection models on acoustic and time-series sensor data from rotating machinery, supporting predictive maintenance.

Also: Research Intern at Nanyang Technological University (pre-trained MedBERT, 50k+ downloads) & SSN College · Graduate Teaching Assistant at McMaster (DSP, Image Processing, Electromagnetics II).

Expertise

Engineering LeadershipTeam management (12+ engineers, tech leads & PMs), hiring & performance management, technical mentorship, partnership engineering (IBM, Intel, AMD, HP), product roadmap, cross-functional coordination.
AI/ML Systems & ArchitectureMulti-cloud (AWS, IBM Cloud) plus on-prem and edge inference; event-driven microservices; RAG, multi-agent orchestration (LangGraph), document intelligence, model quantization (GGUF, INT8/FP16), MCP integration.
Production Infrastructure & StackKubernetes, Kafka, FastAPI, OpenTelemetry, Prometheus/Grafana; Python, TypeScript, PyTorch, vLLM, OpenVINO; PostgreSQL, Redis, Milvus.

Selected Publications All on Scholar · 245 citations →

Google Scholar: h-index 7 · i10-index 6

    Patents

    System and method for Fine Tuning of Large Language Models

    US Patent Application · Filed

    B Sathianathan, L Jothivincent, S Tharumarasa, C Vasantharajan

    Apparatuses, Systems, and Methods for Creating Artificial Intelligence Agents and Routes

    US Patent Application · 2025

    B Sathianathan, C Vasantharajan

    Apparatuses, Systems, and Methods for Generating AI Workflows

    US Patent Application · 2025

    B Sathianathan, C Vasantharajan

    Language Independent Textual Extraction

    US App. 18/590,347 · 2024

    B Sathianathan, C Vasantharajan, S Jacob

    Open Source

    MedBERT

    Pre-trained BERT for biomedical Named Entity Recognition.

    HuggingFace · 568,613 downloads

    NERP

    Python package for fine-tuning transformer-based NER models.

    PyPI · 69,769 downloads

    Tamizhi-Net OCR

    LSTM-enhanced Tesseract models for Tamil & Sinhala legacy fonts.

    GitHub · IALP 2022

    TamilEmo

    Fine-grained emotion-detection dataset for Tamil.

    CodaLab · Shared Task

    Recognition

    • 2026CRN AI 100 — Iterate.ai named to the Top 20 Hottest AI Software Companies.
    • 2025Pinnacle AI Product of the Year — awarded to Generate.
    • 2025TMC Generative AI Product of the Year — awarded to Generate.
    • 2022Top 5 Finalist, HPCIC Innovation Challenge — with team Medispeech@NTU.
    • 2023McMaster Graduate Research Scholarship — competitive grant supporting MASc research.
    • 2018National merit scholarships — Suba Pathum, Mahapola, Ceylinco Pranama & Dialog awards for academic excellence.

    Education

    Master of Applied Science (MASc), Electrical & Computer Engineering

    2023 — 2025

    McMaster University, Hamilton, ON

    • Thesis: SciRAG — A Retrieval-Focused Fine-Tuning Strategy for Scientific Documents, advised by Prof. Thia Kirubarajan. RAG framework for scientific PDFs with custom LaTeX-equation parsing and domain-adapted LLM fine-tuning.

    BSc in Engineering (Hons.), Computer Science & Engineering

    2018 — 2023

    University of Moratuwa, Sri Lanka

    • Thesis: Entity Extraction System for Legal Contracts — party extraction with contextualized span representations on a 1,000-contract annotated dataset.