AI Software Engineer in Toronto

Culture We treat your time as the scarce asset that it is. We respect people being heads-down and having the time to get into a flow. Distractions and context switching are things we work to minimize. We have meetings for specific purposes, and leverage async channels when appropriate. Mash has an open and direct culture. We believe that not making a decision is a decision - and a terrible one at that. We have strong conviction and a bias for action. We can agree to disagree, and move forward after reviewing the details. As we learn, we will reevaluate, iterate and push forward.

About Mash We're a dynamic team with experience at startups, and scaling products at big tech including Google, Amazon, Ada, D2L, CourseHero, Rakuten, Dayforce, PagerDuty, ROSS Intelligence & Lazer. We have top AI staff, including from programs such as MScAC, and advisors that have built startups and led some of the largest organizations in big-tech over the past 15 years. We don't use any seniority titles, and plan not to for as long as feasibly possible.

Role & Responsibilities As an AI-focused software engineer on the Mash team, you will:

• Ship core AI features within the Mash platform.
• Architect LLM-powered solutions and manage related infrastructure.
• Apply information retrieval techniques for search and retrieval-augmented generation (RAG), working with embedding models and vector databases.
• Develop processes and tooling for measuring and iterating on LLM output quality, leveraging eval frameworks and observability platforms.
• Create tooling to collect and leverage internal datasets for evaluation, few-shot prompting, fine-tuning and more.
• Develop bespoke ML/NLP models, when appropriate.
• Build production data pipelines to ingest knowledge sources.
• Stay up to date on cutting-edge research in the ML/NLP space and identify opportunities to apply new techniques within the Mash platform.
• Work closely with a cross-functional team (product, design, engineering).
• Discover and understand user needs.

Our Stack

• AI: Always iterating and changing - recently - Python, Agent Harnesses (e.g., home-brewed, open-source), Inference across open-weight and frontier LLM models, NLP models, speech-to-text, and lots more.
• Application: Python (FastAPI, Pydantic, SQLAlchemy), TypeScript, React Native with Expo, Next.js.
• Data: Postgres, Cloud Storage, Prefect.
• Infrastructure: GCP (e.g., Cloud Run, Cloud SQL, Pub/Sub, Eventarc), Terraform, Docker, Dagger, Kubernetes, Tilt, OpenTelemetry.

Skill Set

• 4+ years of building production software at scale with a multi-year focus on AI.
• Proficiency in 1+ programming languages, such as Python, Go, Rust, C++, or Java.
• You've shipped LLM-powered features to real users: agents and tool calling, context design, retrieval and structured outputs.
• You've built evals, datasets or regression checks that drive meaningful and provable improvement in a customer/user experience.
• Experience with event-driven systems, building cloud-native software (with cloud providers such as Google Cloud Platform, Amazon Web Services, Microsoft Azure, etc.), working with data (both structured and unstructured) and developing data pipelines.
• Strong understanding of backend system design and scalable architectures.
• You use AI coding agents as a normal part of how you build.
• Bonus: agent harnesses, structured interactions, speech or audio pipelines, entity resolution, sandboxing and isolation, privacy-preserving ML.

Cultural Fit We're looking for an individual who:

• Responds well to open-ended problems and thrives on autonomy - is a self-starter and problem solver.
• Thrives in collaborative environments and is laser-focused on team goals.
• Has great communication skills.
• Has a passion for lifelong learning.
• Has strong user empathy, and a user experience first mindset.