About Me
Hi, I'm Anojan.
I'm a Machine Learning Engineer at the Vector Institute, where I build and ship production-grade agentic AI systems for healthcare. My work focuses on designing AI assistants that support real clinical workflows through multi-agent orchestration, LLM evaluation, validation, and reliable deployment.
One of my recent projects involved building a healthcare scheduling assistant that converts natural-language scheduling requests into structured, validated scheduling inputs.
Previously, I worked across applied AI, computational biology, IoT analytics, and medical imaging, building models and ML systems that solve real-world problems. I enjoy working at the intersection of machine learning, software engineering, and product-focused AI.
Outside of work, I enjoy exercising, jogging, playing soccer, hiking, and reading or writing technical blogs.
“If you can dream it, you can do it.”— Walt Disney
Education
Master of Science in Computer Science
Artificial Intelligence • GPA: 3.9/4.0
Western University
London, Ontario
Bachelor of Science (Honours) in Electrical and Electronic Engineering
Sri Lanka Institute of Information Technology
Colombo, Sri Lanka
Work Experience
Vector Institute
Machine Learning Associate
- Building a production healthcare scheduling assistant with Petal Solutions Inc. that converts physician availability and preferences into validated scheduling inputs.
- Designed ReAct-style agent workflows with LangGraph, LangChain, and Azure OpenAI for multi-turn clinical scheduling conversations.
- Added validation, guardrails, and human review loops to improve reliability before downstream scheduling use.
Transpots
Applied AI Developer
- Built agentic document workflows for invoice factoring to extract, validate, and process financial documents.
- Designed LangChain and LangGraph pipelines with GPT-4o, rule-based validation, and human-in-the-loop checks.
- Deployed FastAPI microservices for production extraction, validation, and approval workflows.
Kaidu.ai
Data Science and Engineering Intern
- Built an IoT motion classification system using Bluetooth RSSI data to detect stationary and mobile activity.
- Improved detection reliability with signal preprocessing techniques, including Dynamic Fourier smoothing.
- Used MLflow and Optuna to track experiments, tune models, and support production-ready iteration.
Intelligence Data Science Lab, Western University
Research Associate
- Conducted AI-driven computational biology research for genomic sequence analysis and aptamer discovery.
- Fine-tuned Llama 3.2 and Mistral 7B models for miRNA-target prediction and domain-specific biological sequence tasks.
- Built scalable HT-SELEX data pipelines for preprocessing, feature extraction, and model evaluation.
HeHealth
Associate AI/ML Engineer
- Developed ResNet and VGG16 medical imaging models for clinical image classification.
- Improved model robustness with GAN-based augmentation techniques, including SinGAN and ConSinGAN.
- Integrated Grad-CAM explainability into clinical dashboards to support model interpretation.
SenzMate AIoT Intelligence
Associate ML Engineer
- Built ML pipelines for IoT anomaly detection and predictive analytics.
- Used PySpark, Dask, TensorFlow, and LSTMs to process large-scale time-series sensor data.
- Supported deployment of real-time predictive models for industrial IoT use cases.
Technical Skills
Core technologies and areas I work in across AI, ML, and software development
Programming Languages
GenAI & LLMs
Machine Learning
Natural Language Processing
Responsible AI
Cloud Platforms — AWS & Azure
MLOps & Infrastructure
Data Science & Analytics
Development Tools
Contact Me
Feel free to reach out if you'd like to discuss opportunities, have a quick coffee chat, or just say hi! I'm always looking forward to meeting new people.