Forward Deployed Engineer, AI Delivery

GuruLink; Toronto; On-site; CAD 110,000 - 170,000; Full time

Job summary

GuruLink is seeking an engineer to contribute to a Toronto-based investigative analytics platform used by law enforcement and forensic teams. You will ship AI-enabled software in client environments, building the harness around models with prompts, interfaces, memory, and observability.

The role requires rapid prototyping and turning prototypes into secure, maintainable production code. Expect to work closely with end users, handle complex data pipelines, and navigate trade-offs between

Qualifications

• Shipping production software with AI-enabled components.
• Comfort with TS/JS and Python in production contexts.
• Knowledge of agent loops and multi-step orchestration.
• Experience building AI evaluation frameworks.
• Strong grasp of APIs, distributed systems, and cloud infra.
• Able to turn ambiguity into prototypes and scale.
• Sound judgment on capability, latency, cost, and security.
• Direct collaboration with clients and stakeholders.
• Toronto-based with willingness to travel.
• Background check clearance required.

Responsibilities

• Join a pod rebuilding an investigative analysis platform for law enforcement and forensic teams.
• Develop an agentic layer for data ingestion, analysis, and evidence generation.
• Create prompts, tool interfaces, memory, and observability for the system.
• Champion rapid prototyping, then secure and maintain production-grade code.
• Address failures like hallucination, prompt injection, and unsafe tool calls.

Skills

Production software

TypeScript/JavaScript

Python

Agent loops

AI evaluations

APIs and distributed systems

Problem to prototype-to-prod

Trade-off judgment

Client/cross-functional collaboration

Written and verbal communication

Toronto-based

Background check

Job description

Location: Toronto, Ontario

Forward deployed means what it sounds like. You work inside the client's environment, on their problems, next to their people, and you are measured on software that ships rather than recommendations that circulate.

You will join a pod rebuilding an investigative analysis platform used by law enforcement and forensic teams, adding an agentic layer that ingests device data, analyses it, and produces conclusions that have to survive scrutiny in a formal setting. Your work is the harness around the models: prompts, tool interfaces, context and memory, permission boundaries, observability, and the evaluations that prove the whole thing behaves the way it claims to.

The pace is quick. You take an ambiguous problem, put a working prototype in front of someone within days, then turn it into something secure and maintainable that a colleague can pick up. You will also chase the ugly failures that come with this territory, hallucination, runaway loops, prompt injection, unsafe tool calls, and build the fallbacks that contain them before they reach a case file.

This suits an engineer who wants range rather than a narrow lane. You will touch full stack product work, data pipelines, agent orchestration and infrastructure in the same month, and you will sit close enough to the end user to see what your decisions actually cost them. Patterns you work out here get reused across the firm, so good thinking travels.

Toronto based and in the office. Elsewhere on the East Coast can work, but a daily 8am to 10am Eastern overlap with the client team is required, which rules out the West Coast.

Must Have Skills

• Professional experience shipping production software, with AI-enabled systems somewhere in the mix.
• Strong TypeScript or JavaScript, and genuine comfort in Python.
• Working knowledge of agent loops and multi-step orchestration.
• Experience building evaluations for AI applications.
• Solid grounding in APIs, distributed systems, databases and cloud infrastructure.
• Able to go from ambiguous problem to prototype to production without needing the path drawn for you.
• Sound judgment on the trade-offs between capability, reliability, latency, cost and security.
• Direct client or cross-functional collaboration, since you will not be shielded from stakeholders.
• Clear written and verbal communication.
• Toronto based with occasional travel.
• Must clear a background check.

Nice to Have Skills

• Agent frameworks, workflow engines or model orchestration systems.
• RAG, embeddings, vector search, fine-tuning.
• AI observability, tracing and token or cost analysis.
• AWS, Azure or GCP, containers, Kubernetes, CI/CD.
• Open-source contributions to AI infrastructure.
• Regulated or high-assurance environments.
• Consulting or solutions-engineering background.
• A track record carrying products from discovery through to launch.
• Existing security clearance, or TribalScale experience within the last three months, either of which removes weeks from onboarding.