FAQ
Frequently asked.
Working with KUMO.
Common questions about scope, ownership, production safeguards, delivery model, and where custom work makes sense.
How long does an engagement take? +
Most production engagements run 12-16 weeks. Larger multi-feature builds 4-6 months. We work in milestones, you see progress every 2 weeks, with a working artifact at the end of each.
How do you scope an engagement? +
Discovery comes first: 2-3 weeks where we map workflows, audit data, pressure-test the use case, and align on the business metric this is supposed to move. You get a written scope, a risk register, and a milestone plan before any production build begins.
Should I go custom or buy SaaS? +
Often SaaS. We start every conversation by mapping your use case against existing vendor AI. If a SaaS tool solves 80% of your problem, we'll tell you to buy that. Custom only makes sense when vendor AI can't reach your data, workflows, or compliance environment.
How do you prevent hallucinations in production? +
Every system ships with an eval harness, regression tests, structured output guardrails, source-grounded retrieval where applicable, and a confidence threshold below which the model defers to a human. Accuracy and drift are monitored continuously, we don't ship what we can't measure.
Who owns the code, prompts, models, and data we produce? +
You do. Full IP transfer is the default: code, prompts, fine-tuned weights, datasets, infrastructure-as-code, documentation. Repos are yours from day one. No back doors, no license-locked components.
Will we be locked into a specific model or cloud? +
No. We build behind a model-abstraction layer so you can swap providers (OpenAI, Anthropic, open-source, fine-tuned) without rewriting application logic. Deployment is cloud-agnostic: AWS, GCP, Azure, regional, or on-prem.
Who actually does the work, senior engineers or juniors? +
Senior engineers from start to delivery. The team has 15+ years building production software, including the founding engineering behind Volopay (YC S20). No junior hand-offs after the pitch.
Are you staff augmentation? +
No. Project model by default: scoped deliverables, weekly syncs, close Slack collaboration. Engineers don't report to your managers day-to-day. For clients who need closer embedded collaboration, we adapt the model.
Do you sign NDAs and respect IP ownership? +
Yes. Mutual NDAs and DPAs signed before discovery. All code and IP belong to you on milestone payment. Cross-border IP-transfer clauses are routine, we work across 11 countries.
Can we work alongside our existing team? +
Yes, common pattern. Your team owns product. We own AI engineering or the new build. Weekly syncs, shared backlog, code reviews with your leads, clean handoff at the end with documentation and runbooks.
Where do you keep humans in the loop? +
Wherever the cost of being wrong is higher than the cost of being slow: clinical, financial, legal, customer-facing. Confidence scores route uncertain cases to reviewers, and we instrument those reviews so the model learns from them.
How do you handle data privacy and compliance? +
We implement to SOC 2, HIPAA, GDPR, PCI controls when your engagement requires. We don't hold certifications ourselves, your compliance team owns those. PII masking, encryption, on-prem and confidential-compute deployment available.
Do you only do AI work, or traditional software too? +
Both. Most projects today have AI integrated. We also build standard webapps, mobile apps, and internal tools without AI when that's what the problem actually needs.
What happens after launch? +
Launch is a milestone, not the finish line. Every system ships with monitoring, alerting, rollback paths, and a retraining cadence. You can hand it to your team or keep us on a monthly retainer for monitoring, eval refresh, and incremental improvements.
Where are you based? +
Global team operating primarily from India with active engagements across the US, UK, Europe, the Middle East, and Asia, 11 countries to date. We align working hours to your team's core overlap.
Got a different question?
We'll listen first, ask the right questions, and follow up with a clear proposal.