FAQ
Questions from R&D and automation leads.
Is RoboSparkLab an Apache Spark / data platform, an electrical or science lab, or a consumer-robot toy brand — and do you guarantee results?
No. We are an AI robotics R&D and prototyping lab in Vancouver. The .life TLD is branding only. Spark means the spark of a new robot capability — a prototype coming to life — not Apache Spark, electrical sparks or spark plugs. Lab means a robotics engineering studio, not a chemistry or electrical teaching laboratory. We prototype perception, learning and manipulation, simulate before hardware trials, and report honestly on the sim-to-real gap. Artificial intelligence is probabilistic; the physical world is unforgiving. We test with human operators in the loop and do not guarantee uptime, throughput, zero defects or any specific outcome. We do not build autonomous weapons or unlawful surveillance tools, and we are not a robot store or toy brand.
Do you work on fixed projects or retainers?
Both. Narrow feasibility questions — can this perception stack hit your precision target — often fit a fixed-scope pilot with defined deliverables and evaluation dates. Reinforcement learning, imitation learning and multi-quarter autonomy research usually run on monthly retainers with allocated sim hours and a reserved hardware window in our Mount Pleasant lab. We quote in CAD after a scoping call; indicative ranges appear on our services page.
What budgets should we plan for?
Perception pilots often land between C$18,000 and C$35,000. Manipulation and grasping research typically runs C$22,000–C$48,000. Learning-heavy programmes start around C$45,000 for a multi-month retainer. Digital-twin foundations can begin near C$15,000. Final pricing depends on site visits, sensor hardware you supply versus we procure, and safety documentation depth. We do not ask for full payment before a written scope exists.
How long until we see hardware results?
Most pilots produce a first hardware trial within four to eight weeks if parts and access are ready. Learning engagements need longer sim iteration before unattended runs are even discussed — often ten to sixteen weeks before a meaningful policy demo. We publish a timeline with explicit evaluation gates rather than promising a launch date.
Which platforms and hardware do you support?
We commonly work in ROS 2 with collaborative robot arms, parallel-jaw and vacuum tooling, RGB-D and LiDAR sensors, and autonomous mobile robot platforms supplied by you or rented for the engagement. Edge inference targets are agreed up front — Jetson-class devices, industrial PCs or your existing controllers. We integrate with your stacks where practical; we are not a distributor and do not resell hardware.
How do functional safety and risk assessment fit in?
Every hardware session includes guarding, e-stop verification and operator briefing. We document risk assessment assumptions, reference ISO 10218 and ISO/TS 15066 concepts for collaborative workflows, and note CSA-relevant expectations for Canadian deployments. We support your safety review; we do not replace your certifying engineer or sign off on your production line.
Who owns code, models and IP?
Default terms assign project-specific deliverables to you upon final payment, with RoboSparkLab retaining pre-existing tools and generic libraries. Third-party open-source licences flow through unchanged. Custom terms are available for joint research or shared publication — negotiated before work starts.
How is personal and operational data handled?
Contact form data is used to respond to your enquiry under PIPEDA and British Columbia PIPA where applicable. Project imagery or operator video used for imitation learning is governed by your engagement contract — typically with consent, retention limits and secure storage in Canada or approved subprocessors. See our Privacy Policy.
How do you manage the sim-to-real gap?
We log sim metrics and hardware metrics side by side, publish failure galleries, and stop scope creep when transfer stalls. Domain randomization, teleoperation fallbacks and operator-supervised trials are tools — not magic. If a policy is not ready for your floor, we say so in writing.
What do you not do?
We do not guarantee uptime, throughput or zero defects. We do not sell consumer robots or run a gadget shop. We do not offer get-rich-with-AI courses, robo-advisor products, robocalling tools or self-serve SaaS subscriptions. We do not build autonomous weapons or unlawful surveillance systems. We do not present illustrative metrics as promises of future performance.
Still unsure if we are the right lab?
Send a short note with your capability question. We will tell you honestly if a prototype engagement fits — or if you need a deployment integrator instead.
Our robotics R&D lab prototypes AI-powered perception, learning and manipulation for organizations, with qualified engineers and human operators in the loop. Robotic and AI systems are probabilistic; they act in an unforgiving physical world and can err. Nothing is fully autonomous or perfectly safe. Every hardware test requires site-specific risk assessment, functional-safety measures, guarding and trained operators, and compliance with applicable standards (ISO 10218 / ISO/TS 15066, CSA) and local regulations. We are honest about the sim-to-real gap — simulation success does not guarantee field performance. We do not guarantee uptime, throughput, cycle time, zero defects, cost savings or any specific outcome. We do not build autonomous weapons, lethal systems, or tools for unlawful surveillance. Samples and figures reflect past illustrative work, not promises of future performance. This is a professional robotics-engineering services firm — not engineering, legal or safety-certification advice — and we do not buy or sell personal data.