Greyfeathers StudiosAI-native product studio · Coimbatore, India

100+ products built.
A few of them already
on the road.

We build full-stack products, machine-learning systems and enterprise automation, and we have done it for vehicles, fleets and drivers too. This is a short introduction: who we are, what we have built that is close to your world, and where we might build something together.

Prepared forChristian Fuest · Valtech Mobility
Introduced byPawan · Greyfeathers
DateNovember 2026
→ or space to begin · every project has a live demo you can try
01 · Who we are

Greyfeathers is led by Anandraj and Ruben, with an AI-native engineering team that builds, ships and runs products.

A

Anandraj

CEO, Greyfeathers Studios
  • Built data platforms at Samsung at trillion-byte scale
  • Built location triangulation from movement signals, used to model households and audiences for advertising
  • Worked with Google and McKinsey
  • Leads the studio and its enterprise clients
RS

Ruben Savio

CTO, Greyfeathers Studios · former co-founder & CTO, Baaz
  • Co-founded Baaz, India's first live video shopping platform, and ran its technology
  • Built the geofenced cab-ad network with Uber India
  • Built software that connects head-up displays to cars, and apps for smart glasses and wearables
  • Mechanical engineer; R&D at James Hardie, Chicago; thesis filed for patent by CommScope
Together
0
products built
0
AI applications in production or POC
0
viewers on a single live stream we served
0
buses run on our fleet platform
0
students moved by that fleet every day
0
repositories behind our AI review gate
0
surfaces we ship software on, listed below
0
people in our AI-native team, by choice
Where our software runs
WebAndroidiOSDesktopCar head-up displays (HUD)In-vehicle tabletsSmart glasses & wearablesARBrowser extensionsPhone lines (voice AI)Camera & dashcam feedsLive video streamsSDKs & AI-agent tools (MCP)
What we do
Product designResearch, UX and UI, design systems and clickable prototypes
Full-stack productsWeb, Android, iOS and backend, from first version to scale
Machine learningVision, language, retrieval and voice, in production
AlgorithmsRouting, auctions, optimisation, location intelligence
DevOps & cloudInfrastructure as code, CI/CD, monitoring, cost control
Data platformsPipelines, analytics and dashboards at large scale
Enterprise automationPaper and manual workflows turned into systems
Internal productsTools we build and run for ourselves and our clients
VW
Volkswagen
For a Volkswagen partner in India
Proof of concept

An AI owner's manual for the Volkswagen Taigun and Polo GT.

Problem
Owners dug through a printed manual or called customer support for everyday questions about their car.
What we built
  • A mobile app with an AI engine that searches the manuals for Taigun and Polo GT, every variant and model year
  • Answers come only from the manual of the car the owner bought, with the page images and steps
  • Photograph a warning light and it is identified from that model's manual
  • Manuals in German, English and Hindi; new models and updates added without a rebuild
3languages, every variant and year, and each answer cited to the manual page. Demonstrated as a working app.
owner assistant · answers from your car's manual
Try itIllustrative
My Taigun2024 · manual loaded

Your car

Language

What happens

1 · Question or photoTyped, spoken or a snap of the dashboard
2 · Identify the carOwner's model, variant and year
3 · Search that manual onlyIn the owner's language
4 · Answer with the pageText, steps and images, cited
Uber
Uber India
Built and run in partnership
Ad tech · live

Geofenced ads on screens inside Uber cabs.

Idea
A passenger in a cab is minutes away from the businesses they are driving past. Show them the right one, at the right street.
How it worked
  • A tablet in each cab connected to our ad platform, with the cab's live location
  • Advertisers drew virtual geofences (a mall, a neighbourhood, a stretch of road) and bid for them
  • When a cab entered a fence, the ad played; where fences overlapped, the highest bid won
Advertisers
Textile brandsCafésShopping mallsHotelsEvents
geofence auction · live cabs
Drag the fences · change bidsIllustrative

Advertisers · bid per play

Ad plays0

Auction

23
Education group, Tamil Nadu
23 institutions · day-scholar transport
Fleet platform

Running ~300 buses for ~20,000 students, every day.

One system replaced paper registers, route PDFs and phone calls, for everyone who keeps the fleet moving.

Live operationsGPS on existing trackers, actual vs scheduled at every stop, routes merged or split day by day
PeopleDrivers, cleaners and bus in-charges assigned per trip; standby drivers with a valid licence on call
Breakdowns & incidentsNearest standby bus dispatched; every incident logged and followed to closure
WorkshopMaintenance in rotation, one bus in today and the next tomorrow; job cards, logs, engines and parts traced
FuelIn-house fuel station, every fill against the odometer, mileage alerts per bus
DocumentsInsurance, permits, fitness and licences tracked, with renewals flagged 30 days ahead
fleet ops · morning run
Break a busSimulation

Live

Clock06:30
Buses running0
Students on board0
Reached campus0
Standby buses6
Workshop todayA214

Alerts

In-car voice co-pilot
Built by our team
In-car AI · voice

A voice co-pilot that plans the whole trip, then keeps re-planning it on the road.

Problem
A drive with errands means juggling maps, reviews and traffic on a phone, which is exactly what a driver should not be doing.
What we built
  • Understands spoken requests such as "home, with the groceries on the way" or "A to B with three coffee stops"
  • Compares routes and well-reviewed stops, and picks the best combination
  • Hands navigation to Google Maps
  • Re-checks the route every 5 minutes against live, historical and forecast traffic, and re-routes and re-picks stops without the driver asking
0taps needed from the driver once the trip is planned.
voice co-pilot · in-car
Try itIllustrative

Driver said

Plan

Route watch

Dashcam Vision Engine
Built by our team
Vision AI · dashcam

Dashcam footage that nobody watches, turned into road intelligence.

Problem
Fleets record hours of road video every day. Almost none of it is ever looked at, so nothing in it reaches a decision.
What we built
  • Ingests dashcam and IP-camera video, live or uploaded
  • Segments it into GPS-tagged events: vehicles, people, congestion, potholes, cracks and speed bumps
  • Produces an annotated video, a per-second report and a dashboard
  • Feeds a growing library of segmented road video for training and analysis
1 sresolution in every report: what was on the road, where, and how sure the model was.
dashcam vision engine · cam 01
Move the thresholdSimulation

Detected

0vehicles
0people
0road damage
0events logged

Event stream

Confidence ≥ 0.50
Smart-glasses activity app
Built by our team
Wearables · AR

A sports watch, moved into your line of sight.

Problem
Walkers, runners and hikers keep glancing down at a watch or phone, which breaks their stride and takes their eyes off the path.
What we built
  • Software for smart glasses that tracks the activity live: speed, pace, distance and duration
  • Elevation and total climb along the way
  • A live map of the route shown on the glasses
  • Modes for walking, jogging and running, built for outdoor activity
  • A layout made for a small display at the edge of your vision: glanceable, never in the way
1glance for everything: speed, time, distance, climb and map, without looking down.
smart glasses · through the lens
Try itIllustrative
08 · Across domains

What else we have built.

Beyond the road: the same team, shipping AI products and operations software across domains. Pick a category.

Consumer · social

Hyperlocal video Q&A

Ask a question as a short video and get answers from people in the same place — a location-first take on Q&A, built to rival the text sites.

Built by our team.
Proptech · platform

Property sales & marketing platform

Two years building the system a property developer runs launches on: project microsites, lead capture and CRM, listings and inventory, site-visit scheduling and sales analytics, across web and mobile.

Two years, one sales engine for builders.
Media · live commerce

Baaz

India's first live video shopping platform — real-time shoppable streams that scaled to a million concurrent viewers.

Built by our team.
QA · AI

Testora

Upload manual test cases and an AI agent runs each one in a real browser through Playwright — with screenshots, traces and console evidence — or run scripted, no-AI checks. A test passes only when deterministic browser assertions confirm it.

Our own product · AI browser testing.
Dev tools · AI

AI usage tracker

Measures how a team actually uses its AI coding tools — tokens per person, per tool, per project — and tells each developer the one habit costing them the most.

Coaching "vibe coders" with real data.
Platform · AI + legacy

Legacy event-integration platform

Lifts business events out of legacy COBOL, Java and C# systems with a single emit call — and an AI that reads the repos, finds the logic and opens the pull requests.

Legacy systems, made event-driven.
Agents · engineering

AI code-review gate

Every pull request checked against one engineering standard, editable by anyone, with no deploy needed to change it.

Running across 80+ repositories.
Agents · QA

Droppnotes

Reviewers pin feedback on live software; an AI agent reads the queue with full context and fixes the issue directly.

Our own product, used on client work.
Agents · hiring

HR agent

Turns résumés and GitHub profiles into scored evaluations, generates assessment tasks and reviews the submitted code.

Hiring as a structured-data problem.
Agents · delivery

AI-native delivery

Our own pipeline: design to working screens, parallel frontend and backend, AI tests and AI review.

How we ship.
Document AI · energy

Welding-spec reviewer · SLB

Reads scanned, bilingual welding specs into ~60 parameters and checks each against 26 encoded DNV rules, citing the clause.

A one-day review becomes an hour.
Vision AI · construction

Drawings → quotation

Reads construction drawings into quantities, then prices them with deterministic rate tables into an auditable bill of quantities.

From ~16 hours of take-off to under one.
Document AI · education

CampusVault

Unified document intelligence across multi-campus institutions: find, read and act on records.

Documents, made searchable.
Voice AI

After-hours voice receptionist

Answers calls in natural speech, works out the request, checks a real calendar and books it. Anything unusual goes to a person with the transcript.

Calls that used to ring out become bookings.
Voice AI · healthcare

Ava — AI dental front desk

An AI front desk for dental clinics: takes calls, answers patients and books appointments around the clock.

A front desk that never closes.
Enterprise · platform

SmartOpz

Workplace and operations management for enterprises, run from one system.

Enterprise workplace management.
Education · platform

Cognisho

Runs a whole campus from one screen — operations, people and records in a single system.

A campus, on one screen.
Productivity · delivery

Klarytee

Flags the project that's slipping before it's obviously late, from the signals teams already produce.

See the delay coming.
Sports · operations

Sportiva

College sports run end to end: ground and venue booking, tournament eligibility, team selection, student performance and travel allowances.

A PE department, digitised.
Vertical SaaS · stone

StoneDesk

A sales and operations app for stone businesses — quotes, inventory and jobs in one place.

Industry-specific ops.
09 · Exploring synergies

Where we think our work meets yours.

This is not a proposal. It is a map of how what we've already built lines up with Valtech Mobility's work. If there is a genuine need on your side, we would like to explore building it with you.

A first conversation

30 minutes to get to know each other: you tell us what your teams are working on, we tell you more about how we build. If there is a fit, we take it from there.

ContactPawan, Greyfeathers Studios Germany+49 177 146 2170 Emailpawan@greyfeathers.io Webgreyfeathers.io LinkedInlinkedin.com/company/greyfeathers-pvt-ltd
capability map · drag the nodes
Click a node
Valtech Mobility focusBuilt by our team
Click any node to see how it connects. Drag nodes to rearrange.
Greyfeathers Studios
For Valtech Mobility