My first firmware demonstration worked beautifully until a colleague switched the device off at the wrong moment. I had tested the ordinary sequence and assumed the next instruction would always get its turn. The resulting fault was recoverable on our bench, but the customer would not have a developer sitting beside every installed unit. That was an early lesson in the difference between demonstrating a feature and designing a product's behaviour when its environment refuses to cooperate. It still shapes the questions I ask as CTO.
03 / Record identity across intermittent connections
04 / Whole-device optimisation trade-offs
Engineers led
154
Through five heads in a 520-person company
Annual engineering expense
₹36.8 crore
Separate from ₹7.2 crore laboratory/equipment programme
Embedded engineering experience
26 years
July 2000 to October 2026
Build for the interruption, not just the demonstration
I am chief technology officer of Pelvori Embedded Systems in Nagpur. We write firmware and communications software for industrial equipment makers, including data recorders and diagnostic accessories. I lead 154 engineers through five heads, with ₹36.8 crore annual engineering expense and a separate ₹7.2 crore laboratory and equipment programme. Our clients manufacture the devices and retain final product acceptance. My team must explain what our software has been tested to do and what the result does not establish.
I began in 2000 programming small control-device applications, then moved into firmware services and embedded architecture. The most useful lessons came from conditions that a normal demonstration did not include: an interrupted write, a malformed message or a repeated transmission after a connection returned. I became Pelvori's head of engineering in 2016 and CTO in April 2022. The role now requires me to make the cost of those conditions understandable to commercial colleagues as well as to engineers.
The decisions on this site include accepting extra device memory to make an interrupted update recoverable, widening a message-parser test beyond the examples used to repair it, preserving record identity across an unreliable connection and declining a compression setting that saved transmission volume at an unacceptable processing cost. Each decision changed a delivery conversation. I want an engineering team that can explain the specific evidence behind its confidence without converting a successful test into a universal assurance.
Interrupted updates became a design question
Seventeen of 200 scheduled power-interruption trials originally required bench recovery. A ₹48 lakh redesign produced no bench-recovery cases in the repeated 200-trial matrix. Extra memory added ₹180 to the client's per-device component estimate; field acceptance remained with the manufacturer.
A second message set challenged the repair
A parser failed on 84 of 12,000 malformed laboratory messages. The correction passed that set, but an independent 4,000-message holdout exposed two further failure cases. We repaired those before the client's acceptance trial rather than treating the first clean rerun as complete evidence.
A smaller payload was not the whole result
A compression candidate reduced transmitted bytes by 46% on the agreed sample but raised device processing time by 71%. We selected a lighter option with 27% fewer bytes and 12% additional processing time, after the client reviewed its constraints.
Current focus
I am improving how project teams preserve the conditions of a test when its result reaches a release paper or customer conversation. A number should retain its workload, device configuration and remaining uncertainty, so that an engineering observation does not become a product promise through abbreviated reporting.
Conversations I welcome
I would consider a first external board role in a specialist engineering-services or industrial-software business, with a contribution to product-risk and technology-investment discussions. I have provisionally allowed eighteen days annually, subject to employer agreement, induction needs and a company-specific review of relationships and eligibility.
Beyond the page
The record behind the perspective.
Formation
BE in Electronics EngineeringNarvethi Technical Institute · 2000
Arman brought the extra memory cost and the recovery test into the same decision. He gave us a commercial explanation for the redesign without pretending that a clean laboratory result approved the whole customer's product.
Nadia RelshaniFictional CEO and current reporting manager, Pelvori Embedded SystemsThe full record ↗
Contact
I would consider a first external board role in a specialist engineering-services or industrial-software business, with a contribution to product-risk and technology-investment discussions. I have provisionally allowed eighteen days annually, subject to employer agreement, induction needs and a company-specific review of relationships and eligibility.
