AI agentsbuilt forproduction.
We design, build and hand over AI agents, agentic workflows and the backend and DevOps systems they run on. Guardrails, evals and observability are part of the build, not an add-on.
02 / CAPABILITIES
The parts that make AI useful.
Custom AI agent development, agentic workflow automation, backend engineering, and the infrastructure to operate them.
AI agents
Purpose-built agents that use your tools and knowledge, know their limits, and can be evaluated in production.
02Agentic workflows
Multi-step operations that connect systems, route decisions, request approvals, and leave a clear audit trail.
03Backend systems
APIs, data models, queues, auth, and observability that keep AI products dependable as usage grows.
04DevOps for AI
Deployment pipelines, runtime observability, secrets, rollbacks, and incident readiness for agents and workflows.
03 / WORK
Products we build and run ourselves.
The same engineering discipline we bring to client work, running in products under our own name.
Supaintro
Turn a product URL into a launch film with brand extraction, a storyboard, motion graphics, voice, and captions.
Outstation
Travel planning that turns an open-ended idea into a day-by-day itinerary.
OCDify
Local-first file organisation with on-device classification and control.
GymHive
Membership management, WhatsApp engagement, and analytics for fitness centres.
Have a workflow that should run itself?
Let's map what a reliable version would take.
Start a conversation04 / STUDIO
Enterprise experience behind the studio.
Vyan Labs is led by Sourish Sarmah, a backend and applied AI engineer with experience across product and enterprise environments.
Before Vyan Labs, Sourish built production systems for teams in private markets, construction technology, healthcare, and data infrastructure. That experience shapes how we approach reliability, ownership, and handover.
Founder experience / prior rolesAgentic Private Markets
Multi-agent research, realtime voice, and an enterprise Text-to-SQL MCP with security and guardrails.
Tailorbird
RFP and document processing systems, invoice extraction, audit logging, and engineering leadership.
Biofourmis
Medical data sync services, performance work, and backend migration.
Inviz AI / Tata CLiQ
Data platform and ETL work for a Tata CLiQ project while at Inviz AI.
These are Sourish’s prior roles and project experience, not a list of Vyan Labs clients or endorsements.
Read the enterprise Text-to-SQL MCP engineering note05 / APPROACH
A clear path from idea to operated system.
We start with the actual workflow, expose the risks early, and hand over something your team can run.
Map the job
Define the users, decisions, integrations, constraints, and success measure before choosing an architecture.
Make it work
Ship in useful increments with real data, clear approval points, tests, and evaluation criteria.
Keep it running
Instrument the system, document the decisions, and give your team the controls to own it.
06 / QUESTIONS
A few things worth knowing.
The practical details that usually come up in a first conversation.
Do you build prototypes or production systems?
Both. A focused pilot can establish feasibility, but the design includes the path to monitoring, evaluation, recovery, and team ownership.
Can you work with our existing stack?
Yes. We review the current APIs, data sources, security requirements, and deployment environment before proposing a build.
Who will work on the project?
Sourish leads the technical work directly. Any additional specialists are discussed with you as part of the scope.
Where do you work?
Vyan Labs is based in Guwahati, India and works remotely with teams across time zones.
07 / CONTACT
Tell us what needs to work.
A useful first conversation starts with the workflow, constraints, and what success would look like.
Let’s make it real.
Share a brief outline. Sourish will reply directly to discuss scope, feasibility, and a sensible first milestone.
support@vyanlabs.in ↗