Mitacs Accelerate Research Intern (AI/ML) — Self‑Healing Web Automation

About Lodgr

Lodgr is a Canadian travel technology company building a new approach to short-term rentals. Founded to address the growing imbalance between platforms and property owners, we give hosts more control, transparency, and flexibility over how they operate and earn. Built with speed and focus and guided by what matters most to hosts and guests, Lodgr is early in its journey, which means decisions are made quickly, feedback travels fast, and everyone has a direct impact on what we build.

Why You Should Work Here

Youll thrive here if you love building, moving quickly, and turning ideas into action. This is a place for curious, motivated, detail-oriented people who take ownership and enjoy seeing work through, from start to finish. Youll work in a fast-moving environment where decisions are made quickly, accountability is clear, and feedback is open. If youre energized by impact and excited by the opportunity to help shape how a growing company operates, youll find meaningful and rewarding work here.

Summary

We are seeking a Mitacs Research PhD Student to lead the next phase of development for our self-healing browser-extension framework. This role will focus on advancing our proof-of-concept into an efficient, intelligent, and scalable system capable of operating reliably across a wide range of dynamic websites. You will design, implement, and evaluate novel approaches in caching, transfer learning, reinforcement learning, and predictive maintenance to ensure the agent not only reacts to website changes but proactively anticipates and prevents failures.

Day‑to‑Day Responsibilities

• Design experiments; implement prototypes for semantic caching, context-aware repair, RL-based strategy selection, and predictive risk modeling.
• Build data instrumentation: capture execution traces and develop failure taxonomy for RL mechanisms and pattern mining.
• Train and evaluate models using semantic annotations and historical patterns; implement multi-agent specialization and RAG, compare strong baselines.
• Generate metrics tracking performance (latency, cost, success rates) and research insights (failure distribution, healing strategy effectiveness, RL learning curves).
• Write research notes and contribute to Mitacs deliverables; frame novelty in combination of declarative flows + semantic breakdown + agent-driven adaptation.
• Collaborate with engineers and supervisor; present findings in weekly reports and presentations.

Minimum Qualifications

• Education: Enrollment in a Canadian university graduate program (PhD preferred, strong Masters considered) meeting Mitacs eligibility requirements.
• Engineering Foundation: Strong proficiency in Python and modern ML frameworks (PyTorch or TensorFlow/JAX).
• Web Automation: Hands-on experience with web automation tools (Playwright, Puppeteer, or Selenium) and a deep understanding of DOM, CSS selectors, and XPath.
• Applied AI Concepts: Demonstrated foundational knowledge in at least two of the following domains critical to the role:
• Reinforcement Learning (RL) or Contextual Bandits.
• Large Language Models (LLMs) and Prompt Engineering.
• Information Retrieval or Retrieval-Augmented Generation (RAG).
• Research & Methodology: Solid grounding in statistics, experimental design, and algorithmic thinking.
• Communication: Clear written and verbal communication, with a disciplined approach to research documentation and reproducibility.

Nice‑to‑Have Experience

• Specialized AI Tooling: Experience with specific RL toolkits (e.g., RLlib, SB3, offline RL) or vector databases (e.g., FAISS, Milvus, Pinecone).
• Deep Browser Internals: Familiarity with the Chrome DevTools Protocol (CDP), accessibility trees, or front-end observability. Experience specifically building browser extensions is a major plus.
• Production Systems: Experience scaling prototypes using caching/distributed systems (Redis, Kafka), microservices, Docker, or CI/CD pipelines.
• Predictive Analytics: Background in change detection, time‑series modeling, or anomaly detection (highly relevant for predictive risk modeling).
• Mitacs/Industry Experience: Prior experience balancing academic research deliverables with fast-paced industry requirements.

Position Type

• Academic Partner: University of Waterloo
  
• Location: Vancouver, BC / On-site
• Term: 4–8 months (Mitacs Accelerate; possibility of extension)
• Start Date: May 2026 (Spring Term)
• Compensation: Mitacs Accelerate stipend + top‑up per program guidelines

Please note: Candidates must be currently authorized to work in Canada.

We are an equal opportunity employer, and all qualified applicants will receive consideration for employment without regard to race, colour, religion, sex, national origin, disability status, protected veteran status, or any other characteristic protected by law.