Let’s Talk About Your Project
Whether you need an LLM evaluation pilot, a quote for 3D LiDAR annotation, or just want to understand if we are the right fit — reach out. We respond within 24 hours on business days, usually faster.
Two Ways to Begin
Pick whichever is easier. Both lead to the same place — a clear understanding of whether we are the right fit for your pipeline.
Book a Free 15-Minute Consultation
Talk to Keylian directly. Bring your questions about LLM evaluation methodology, 3D annotation requirements, QA processes, or pricing. No sales pitch — just a technical conversation about your specific use case.
Book on Calendly →Send Us a Message
Prefer async? Email us with a description of your project. We will reply within 24 hours with relevant questions or a call invite to move things forward.
Email the Team →From First Contact to Production in About a Week
Here is exactly what the onboarding process looks like after you reach out. No mystery, no weeks of back-and-forth before you see anything real. Click any stage to see what happens.
You Send a Message or Book a Call
Email us or book on Calendly. No lengthy intake form, no “fill out this questionnaire first.” A brief description of your project or a calendar slot — either one works. We will figure out the rest together.
We Reply With a Few Questions or a Call Invite
We confirm we received your message and either ask a few clarifying questions to scope the work accurately, or confirm the calendar slot. If you have sensitive data, we send an NDA at this stage — executed within 24 hours so nothing is delayed waiting on paperwork.
We Scope the Pilot or Project
We review your data, requirements, and timeline together. For LLM evaluation, we align on the rubric and calibration criteria. For 3D LiDAR, we review sample frames and confirm format and object classes. This is where we set the QA standard we will be held to.
The Free Pilot Starts
Raters are assigned and trained on your rubric before touching your data. For LLM eval, calibration tasks run first. For 3D, annotators are briefed on your specific object classes and edge case definitions. Then production begins on your actual data — 200 to 500 items at zero cost.
You Receive the Pilot Output and Accuracy Report
Full QA-verified output delivered in your format alongside an accuracy report covering inter-annotator agreement, flagged disagreements with rationale, and batch-level metrics. This is where you evaluate whether the quality meets your standard.
You Review and We Discuss Any Adjustments
You go through the output. If there are guideline adjustments to make, we make them. If the quality looks good, we scope the production run and confirm pricing. No obligation to proceed — but most pilots do lead to production because you have already seen what we deliver.
Production Runs. Quality Compounds Over Time.
Full production at agreed volume and pricing, locked for the project window. The rubric and guidelines built during the pilot carry forward and deepen with every batch. No renegotiations mid-pipeline, no surprises on the invoice.
Direct Contacts
Three ways to reach us. All go to the same team — use whichever fits the situation.
For project enquiries, pilot requests, partnership discussions, and vendor applications. We respond within 24 hours on business days.
[email protected] →Calendly
Book a 15-minute consultation directly into the calendar. No back-and-forth scheduling. Pick a time that works and we show up prepared.
Schedule a Call →Connect with Keylian directly on LinkedIn. Good for quick questions, introductions, or if you prefer starting a conversation there before moving to email.
Connect on LinkedIn →Location
Based in Nairobi, Kenya. GMT+3 gives us morning overlap with European clients and afternoon overlap with US East Coast teams.
Nairobi, Kenya · GMT+3Response time: We reply to all enquiries within 24 hours on business days. For urgent requests, mention “URGENT” in your subject line and we will prioritise.
What to Include in Your First Message
You do not need all of this to reach out — we will ask what we need. But the more context you share upfront, the faster we can give you an accurate scope and timeline.
Service Type
LLM evaluation, RLHF preference labeling, agent benchmarking, audio CSAT, 3D LiDAR cuboids, sensor fusion, 2D image annotation, or a combination.
Data Format
For LLM eval: prompt/response pairs, audio files, agent traces. For 3D: point cloud format, sensor types. For 2D: image or video, resolution, scene type.
Dataset Size
Approximate number of items — evaluation pairs, frames, sequences, or audio clips. Even a rough estimate helps us scope the work accurately.
Timeline
When you need the pilot or first batch delivered. If there is a hard deadline or a model release driving the schedule, let us know upfront.
Sample Data
A few example items if you can share them. We sign NDAs before any data transfer — just mention it and we will send one immediately.
Output Format
KITTI, nuScenes, Waymo, COCO, YOLO, JSONL, CSV, or a custom schema. If you have an existing pipeline, we match it.
When We Are Available
GMT+3 gives us genuine overlap with European mornings and US East Coast afternoons — not just async delays.
Priority response available for dedicated retainer clients. Project emergencies handled outside office hours on request.
Questions We Get a Lot
How does the free pilot work exactly?
You share your data and rubric (or we scope the rubric together on a call). We annotate 200 to 500 items at zero cost, run them through our full 4-layer QA pipeline, and deliver the output with an accuracy report. No payment information required. Typical turnaround is 48 hours.
Do you sign NDAs before seeing any data?
Yes, always. Send your NDA template to [email protected] and we will have it executed within 24 hours. If you do not have a template, we can send ours. Nothing moves before it is signed.
How quickly can a project actually start?
Free pilots start within 48 to 72 hours after scoping. Full production projects begin 3 to 5 business days after contract signing and data transfer. Rush starts are available with 48-hour advance notice.
Will I be talking to a salesperson or the actual team?
Keylian (Founder and CEO) handles all initial conversations. If the project involves specific technical requirements, Ibrahim (COO and QA Lead) joins the call. You will not be bounced around a sales org.
Do you only do LLM evaluation now?
No. LLM evaluation is our primary service lane but 3D LiDAR annotation, sensor fusion, AV edge cases, robotics 6DoF, and 2D computer vision annotation are all fully active. Same quality infrastructure, same QA pipeline. We just lead with the lane where demand is strongest right now.
What if I need multilingual evaluation capacity?
We have French, Chinese (Mandarin), and German speaking raters available alongside our English-core team. Multilingual capacity is priced at approximately 30 to 60% above English-only rates depending on language and task complexity.
The Pilot Is Free. The Quality Speaks for Itself.
200 to 500 items on your actual data. 48-hour turnaround. Full accuracy report. You see what we deliver before signing anything.