: Staff AI Engineer

Location: Toronto, ON (Scarborough)

Reports To: Director of Technology

Company: ATS Software

Position Overview

We are looking for an experienced and hands-on Staff AI Engineer to serve as technical lead and architect for our AI initiatives.

You will own the architecture behind automated engineering specification matching, PDF and blueprint data extraction, intelligent product cross-referencing, and predictive pricing across the ATS product suite. The central judgment in this role is build versus buy: deciding which capability is best served by a frontier model API, which by a specialized document AI service, which by a self-hosted open-source model, and which should not use AI at all — then owning the cost, accuracy and latency consequences of those calls.

The core ATS application is Node.js. AI services may be written in Node.js or Python, and we expect you to be fluent in both or to become so quickly. You will work closely with Product Management, Data Engineering and Core Platform Engineering to turn domain problems in construction and distribution into systems that hold up in production.

Key Responsibilities1. Architecture & Technical Leadership

- Architect and deploy end-to-end AI pipelines for construction technical documents, submittals, schedules and product cross-referencing.
- Own the build-versus-buy decision per capability: benchmark frontier LLMs (Claude, Gemini, OpenAI), specialized document AI services (AWS Textract, LandingAI) and self-hosted open-source models against accuracy, latency and cost per document, and defend the choice with data.
- Establish the design standards and practices the team works to: retrieval and RAG patterns, prompt and schema conventions, evaluation, production monitoring, and the routing and fallback logic between providers.
- Serve as a senior technical authority across the organization, mentoring engineers and guiding technical decision-making for the data and AI platforms.
- Say no to AI where a deterministic solution is better, and make that case to product and executive stakeholders.

1. Core AI Capabilities

- Document intelligence and extraction: Build extraction pipelines for unstructured technical spec sheets, MEP blueprints, equipment schedules and quotes, combining OCR, multi-modal models and schema-constrained outputs with validation and confidence scoring.
- Product matching and cross-referencing: Build semantic search, embeddings and reranking to map manufacturer products and part numbers automatically, including the domain tuning that generic embeddings do not give you on trade jargon and part number formats.
- Automated CPQ and quoting: Apply retrieval and generative workflows within AutoQuote to cut quote turnaround and improve accuracy for wholesalers and distributors, with human review on anything the system is not confident about.

1. Evaluation & Quality

- Own the evaluation strategy across every AI feature: golden datasets, automated regression suites, and the rule that no prompt, model or retrieval change ships without a run.
- Set the accuracy bar with Product before a feature is committed,



and define what happens below it — confidence thresholds, human review routing, and graceful failure.
- Build the tooling and benchmarks that let the team compare providers and models on ATS’s own documents rather than on public leaderboards.
- Monitor production output quality for drift and failure clusters; drive data governance, quality and privacy standards for all training and evaluation data.

1. Productionization, Performance & Cost

- Integrate AI services into ATS’s cloud microservices architecture (AWS), ensuring high availability, low latency and cost efficiency.
- Own the unit economics: cost per document and per quote, token spend, caching and batching strategy, and the vendor rate limits and failure modes that shape the architecture.
- Implement guardrails, retry and fallback paths, and the continuous learning loops that turn reviewed output back into evaluation data.
- Work with Software and DevOps engineers to implement CI/CD for AI features, with evaluation runs gating deployment.

1. Cross-Functional Collaboration

- Partner with Product Managers and C-level stakeholders to evaluate feasibility, estimate impact and define technical roadmaps for AI features, including setting realistic accuracy expectations before commitments are made.
- Translate what the models can and cannot do into terms the business can plan against.

Qualifications & ExperienceRequired
- Education & Experience: 7+ years of software engineering experience, with 5+ years building and deploying AI/ML systems into production.
- Generative AI & Retrieval: Deep hands-on experience with LLMs, RAG architectures, prompt engineering, structured outputs and tool use, and vector databases (Pinecone, Qdrant, Weaviate, pgvector). Familiarity with LangChain/LlamaIndex, or a clear point of view on when to skip them.
- Model Selection & Economics: A track record of choosing between vendor APIs, specialized document AI services and self-hosted open-source models on evidence, and owning the cost and latency consequences at production volume.
- Evaluation: Has built the evaluation layer, not just the feature — golden datasets, regression suites, accuracy targets, drift monitoring. Should be able to describe how a team under them knew a change was an improvement.
- Design Patterns & API Craft: Fluency with common design patterns and the judgment to know when not to apply them, plus experience refactoring existing REST APIs without breaking their consumers.
- Production MLOps: Proven record of taking AI systems from notebook to scalable, monitored microservices using Docker and AWS.
- Document Processing & Vision: Experience parsing complex unstructured documents (PDFs, CAD/BIM tables, drawings) using OCR,



multi-modal LLMs or layout-aware models.
- Engineering Rigor: Strong computer science fundamentals, system design expertise, clean code practices, and real skill with SQL databases (PostgreSQL, MySQL).
- Languages: Highly skilled in Node.js/TypeScript or Python — not merely familiar.

Nice to Have
- Background or domain knowledge in AEC (Architecture, Engineering, Construction), PropTech, CPQ, supply chain or technical e-commerce.
- PHP and REST API experience — parts of the platform are PHP REST services.
- Deep learning experience (PyTorch or TensorFlow), particularly fine-tuning embedding models or rerankers for domain-specific retrieval.
- Traditional ML methods (XGBoost, gradient boosting, regression) for the pricing and forecasting problems where they beat an LLM.
- Experience building user-facing AI features that require high accuracy and explainability.
- Experience mentoring or leading a small team of engineers.

Compensation $100,000 – $205,000 CAD base per year, plus benefits. Scarborough, ON.

Your Life and Career at ATS:

ATS is dynamic, industrious, creative and collaborative. We believe diverse and inclusive organizations create work environments that are inventive and open-minded, where people spark new ideas and explore alternatives. At ATS, we offer the following to make sure you have a rewarding and enjoyable experience:

· This is a new position

This is an in office position

· Career advancement opportunities

· Benefits package for all eligible full-time employees (including medical, vision and dental).

· A culture that promotes a healthy, fulfilling work-life balance

· Free parking

· Foosball, Ping Pong Table & Basketball net

· Gym facilities

· Epic year-round employee events!

TO APPLY: Please email [email protected] and include your resume and salary expectations. NO PHONE CALLS PLEASE.

To learn more about our company visit our web page: www.atssoftware.com/ and our promotional video: http://youtu.be/MPyk3BdN-8o

Allied Technical Sales Inc. values diversity and is proud to be an Equal Opportunity Employer. We are committed to the principles and practices of employment equity and encourage all qualified individuals, including women, persons with disabilities, visible minorities, and Aboriginal Peoples to apply. Should you be individually selected to participate in an assessment or selection process, accommodations are available upon request in relation to the materials or processes to be used.

Pay: $100,000.00-$205,000.00 per year

Benefits

- Casual dress
- Company events
- Dental care
- Disability insurance
- Life insurance
- On-site gym
- On-site parking
- Paid time off
- RRSP match

Application question(s):
- Do you have at least 7 years of software engineering experience, including 5 or more years building and deploying AI/ML systems into production?
- Are you highly skilled and fluent in both Node.js/TypeScript and Python?
- What is your hands-on experience designing and deploying RAG architectures, vector databases, and MLOps pipelines on AWS?

Work Location: In person

📌 Staff AI Engineer (Scarborough)
🏢 Allied Technical Solutions
📍 Scarborough