Nebulix Labs · Applied AI & software

Turn complex work into software that works for you.

We help businesses and research teams automate demanding workflows, put AI to practical use, and bring new products to life.

Our own productBuilding Vayun Current researchClinical AI research Founder’s experienceProduction systems

In practice

Products, research & engineering.

Our own product, current research, and production systems carried into practice.

01 A product of Nebulix Labs

Vayun

From customer messages to finished bookings.

Customer questions, scheduling, and payment follow-ups can take over the working day. We built Vayun to connect those steps in one AI-assisted workflow.

What we built
A business assistant for customer answers, lead capture, bookings, reminders, and advance payments.
Inside the system
Business context, conversational AI, scheduling, and payment integrations across messaging and web channels.
In practice
Built and deployed as our own product, with development continuing as we shape Vayun.
  • WhatsApp
  • Telegram
  • Website
  • Razorpay
  • UPI
Explore Vayun
Vayun assistantIllustrative conversation · appointment booking
Rohan

Need haircut tomorrow evening. Only Arjun if possible.

Vayun

Arjun is free after 6:30 PM tomorrow. You usually book Haircut + Beard Trim, which takes around 45 minutes. You can come anytime after 6:30. What time works best?

Rohan

Maybe around 7, I might be late.

Vayun

Then 7:15 PM is better. Booked for tomorrow at 7:15 PM with Arjun. Reminder set 2 hours before.Pay ₹499

02 Ongoing research

Clinical AI research

Applied research for complex clinical and scientific workflows.

Current research at ASRAM explores how AI models can be trained, evaluated, and validated for clinical research questions. The work spans diagnostic, imaging, and laboratory-focused studies, with methods and tooling adapted to each research problem.

Focus
Model training and validation, data preparation, evaluation workflows, and research tooling.
Stage
Research in progress. Project methods and results will be shared where publishable.

03 Earlier production engineering

Production systems built for Scotiabank

Before founding Nebulix, Vijay built production automation, analytics, and generative-AI systems through Tech Mahindra for Scotiabank, Canada.

01

Operational analytics & forecasting

Built an analytics dashboard over an ETL pipeline covering thousands of RPA bots, with performance reporting and an ML forecasting model for future workload demand.

Technical depth
ETL pipeline, operational analytics, dashboard design, and machine-learning forecasting.

02

Recurring workflow automation

Designed an asynchronous automation system for a monthly activity that consumed 40–50 hours of distributed manual work.

Outcome
Enabled the recurring process to be completed within one day each month.

03

Conversational access to production data

Built a generative-AI chat interface connected to an existing production database, so users could query live operational data in natural language.

Outcome
Reduced the need to navigate application records individually through the UI.

Founder’s earlier work · Tech Mahindra engagement for Scotiabank, Canada

Vijay Penmetsa speaking at a lectern bearing the ASRAM name.
Vijay speaking at ASRAM.

Meet the founder

Vijay Penmetsa

Founder, Nebulix Labs

Vijay brings together applied AI research and production engineering. His work spans clinical model training and validation, operational software, and real-time inference at the edge. Alongside this, he is building Vayun, Nebulix’s own AI product for business operations.

Earlier, at Tech Mahindra, he built automation, analytics, and generative-AI systems around live production environments. That range shapes how Nebulix works: understand the domain, choose the right methods, evaluate them carefully, and carry the engineering through to a usable system.

Our approach

A clear path to a useful system.

01

Understand the work.

Start with the people, workflow, data, and constraints. Agree on what a useful result looks like.

02

Test the important assumptions.

Build a focused first version. Evaluate it against the task, with clear points for review and human judgment.

03

Carry it through.

Connect the system to the tools it needs, prepare it for deployment, and make the handover clear.

Let’s talk

Book an appointment with us.

Use the Nebulix assistant to choose a time that works for you.

Emailhello@nebulixlabs.com

Phone+91 94909 43338