In manufacture
2 Neural Processing Units
AI models run on the device itself: thousands of multiply-accumulate units work in parallel, with memory right beside them.
Trijal P G
Open to founder’s office, product, hardware & edge-AI roles
I build the company and the board it ships.
I’m Trijal, founder & CEO of Qynetic Labs, an edge-AI hardware startup in Chennai. Hardware, software, security, and the plan that pays for them.
--:--
26public repos
- retail-edge-aiOct ’26
- Igaos-publicSep ’26
- igaos-solver-studioSep ’26
Vayu One
12 layers. 182 parts. Zero unrouted.
- 100 × 70 mm
- 2,038 vias
- 3 mil tracks
A platform, not a cheap board.
Qynetic Labs makes edge-AI computers: AI that runs on the device, not in the cloud. Our own hardware, OS, SDKs and fleet tools, with public-sector channels and local support.
Founded 2024 · Private Limited since June 2026
- India-local trust
- Developer ecosystem
- Own silicon
Four domains, one chip.
Five boards. One software stack.
From a developer board to a 434 TOPS edge server, all running QyneticOS and the same SDK. Tap a board for its specs.
Our own silicon, in three steps.
Board → silicon → platform.
2
In manufacture
Neural Processing Units
An NPU runs AI models on the device itself. Thousands of multiply-accumulate units do the maths in parallel, while on-chip memory keeps weights and data close, so inference is fast and light on power.
Drag to rotate · tap to open
Inside the NPU die
- Design complete
- In manufacture
- Bring-up
1
Manufactured · in the Vayu lineup
Tensor Processing Unit
Built for cost-efficient inference. A TPU is designed around matrix multiplication, the core operation of every neural network: data and weights flow through a grid of processing cells in lockstep, so each value is reused many times instead of being fetched again.
- 197
- TFLOPS · BF16
- 393
- TOPS · INT8
- Inside
- ·
Drag to rotate · tap to open
How the TPU computes
- Design complete
- Manufactured
- In Vayu Pro & Ultra
- Founder & CEO, Hardware LeadQynetic Labs Private Limited · 2024 – Present
- Data Center Systems & IT-OT InternAfcons Infrastructure Ltd. (Shapoorji Pallonji Group) · May 2026 – Aug 2026
- Freelance Web & App DeveloperSelf-employed · 2019 – Present
B.E. Computer Science & Engineering (Cyber Security) · Chennai Institute of Technology · 2028 · IIT Madras B.S. (qualified)
Syncing with GitHub…
–contributions, last 12 months
–active days
–longest streak
–followers
Contributions per month
Languages by code size
It started with a website.
- 2019First paid website
- 2021National runner-up, web design
- 2022National winner, web design
- 2024Started Qynetic Labs
- 2025Built YUNA
- 2026Pvt Ltd, Vayu One, SIH, Afcons
Say hello.
Hiring for a founder’s office, product, hardware or edge-AI role? Building something at the edge? I’d like to hear from you.
Personal: trijalgunaseelan13@gmail.com