A familiar retail label concealed a hardware revision that mattered to software. Compatibility needed evidence about the device, with the right technical owners retaining approval.
Fictional Veznori Sewing Retail, based in Coimbatore
Product-engineering team
64 employees
Eighteen contractor positions are additional; six reporting leaders
FY2026–27 delegated authority
₹26 crore
₹18.2 crore operating expense plus ₹7.8 crore approved product-engineering work
Build for the equipment people already own
I am Leela Qavreti, CTO of the fictional Veznori Sewing Retail in Coimbatore. We sell sewing equipment through forty-four showrooms and develop software that helps customers prepare and use embroidery projects. I grew up in Surat and began writing embedded software for instruments. That work taught me to ask which physical revision a program would meet. Moving into consumer products added another question: what happens when the customer's computer, connection or older machine differs from the equipment on which we made the demonstration?
Veznori employs 860 people and recorded ₹396 crore revenue with ₹31.4 crore operating profit for the year ending March 2026. Founders own 58%, a private investor 32% and an employee holding vehicle 10%; I have no personal holding. My remit covers sixty-four employees and eighteen separate contractor positions through six leaders. FY2026–27 authority comprises ₹18.2 crore operating expense and ₹7.8 crore approved product-engineering work. The CIO retains enterprise systems. I report to CEO Arvind Talqeshi, and engineering quality head Mira Zervathi leads software acceptance with the relevant qualified manufacturers where hardware is involved.
Recent decisions have made those boundaries practical. We stopped using a retail model name as sufficient evidence for a firmware package. We removed an unnecessary cloud-session dependency from opening eligible projects already stored on a customer's computer. We funded preview work against ordinary customer machines instead of reporting the speed of an engineering workstation. A supplier handover then showed why receiving source files was different from being able to build and test the software. Each programme required a specific account of what we could support, what remained outside scope and who retained the authority to approve it.
A retail model name concealed hardware differences
Eight of 100 bench devices did not match the hardware assumptions of the proposed package. We funded explicit compatibility checks and delayed the pilot by seven weeks.
An online session was blocking existing local work
Seventy-four of ninety-six outage reports concerned eligible projects already downloaded. We separated reopening that work from transactions requiring an online check.
A source archive did not yet provide a working build
Eight of eleven supplied components initially built in the agreed environment. Three needed dependency or tooling resolution before the handover could be used.
Current focus
Making supported equipment, offline behaviour and reproducible delivery explicit in the customer-product engineering account.
Conversations I welcome
One first non-executive opportunity after preparation, employer consent and a realistic review of the proposed remit.
Beyond the page
The record behind the perspective.
Formation
Bachelor of Engineering in ElectronicsVelruni Engineering College · 1997
Master of Technology in Embedded Systems, part-timeQezravi Technical Institute · 2006
Leela reported the source handover as a usable build capability with two features still deferred. That distinction let us recognise the result without pretending the transition had completed every release obligation.
Arvind TalqeshiCEO, Veznori Sewing Retail; fictional colleagueThe full record ↗
Contact
One first non-executive opportunity after preparation, employer consent and a realistic review of the proposed remit.