From defense-grade embedded testers to semiconductor equipment, telecom mediation, and automation platforms: more than two decades of engineering reliability into the systems nobody sees.
The best version of hidden-infrastructure engineering disappears: the machine runs, the record is correct, the test catches the issue before anyone outside the room notices.”
Nobody celebrates a billing pipeline that did not corrupt data today. Nobody notices a safety check that quietly prevented a failure before the morning shift began. When the systems Sudhir Kumar Verma builds are working, the public barely registers that they exist, which is, in a sense, the whole point of his career.
Technology is usually introduced to the world through the parts people can see: faster chips, smarter phones, safer trains, automated warehouses, networks that seem to run with effortless precision. The real work of making those things reliable happens underneath the visible layer: in calibration scripts, diagnostic tools, embedded controllers, mediation pipelines, and automation routines that never appear in an advertisement. Verma’s career belongs to that hidden layer. Across a journey that began in 2004, he has built defense-grade embedded test systems, biometric public-benefit terminals, satellite-broadband device managers, carrier-grade telecom mediation platforms, warehouse automation software, and, most recently, diagnostic automation for advanced semiconductor lithography equipment. The industries changed. The engineering pattern did not.
DEFINED BY CONSEQUENCE
The most accurate way to describe his domain is mission-critical software engineering for high-precision semiconductor equipment, industrial automation, and large-scale telecom data platforms: one coherent field rather than a scattered list of employers. Mission-critical, in his usage, does not mean every project was life-or-death. It means the software supported systems where failure carried real consequences: a faulty diagnostic, a disrupted billing pipeline, an unreliable field terminal, a test environment that let a defect slip through. Work at that boundary keeps asking blunt questions. How do you know a multi-million-dollar machine is healthy before a technician touches it? How do you shorten a regression cycle without losing coverage? How do you move billions of telecom records without a single loss or duplicate? How does a field terminal complete a transaction when the network is broken?
“The industries changed; the engineering pattern did not: build the system, validate its behavior, diagnose the failure, and make the result trustworthy.”
FOUNDATIONS: THE HANGAR AND THE COUNTER
The foundation was laid at Applied Electro Magnetics in Noida, where Verma developed embedded testing and simulation software for the Slat/Shock Cone and Air Data Simulator systems used with Indian Air Force Mirage 2000 support equipment. Working in C, Visual Basic, and SQL, he wrote software that simulated complex aircraft signals to exercise onboard avionics behavior, and supported deployment at the customer site. Ground-support equipment is easy to underestimate because it sits behind the aircraft rather than inside it, but it is a pillar of maintenance confidence; it lets a technician feed known inputs into a system, verify the exact response, catch drift, and make a definitive go/no-go decision. It was a formative lesson in treating software as a reliability instrument rather than a feature-delivery tool.
At GISIL Designs, he carried that discipline into public-sector technology: application software for biometric fingerprint scanners, the AT-command logic driving GSM modems, and integration work binding authenticated identities to public-distribution transactions at ration counters, under field conditions where connectivity and power could not be assumed. At Hughes Systique, the work moved into distributed communications infrastructure: device management for satellite-broadband customer-premises equipment and network-operations components, covering remote provisioning, configuration, firmware upgrades, and diagnostics over protocols such as TR-069, SNMP, and HTTP/HTTPS. The common theme was reliability under distribution: operating, monitoring, and diagnosing dispersed devices without a field visit.
THE ERICSSON YEARS
His longest run of domain depth came at Ericsson, on the Revenue Manager, Multi Mediation, and Mediation platforms. Telecom mediation is one of the least visible, most consequential layers of a communications network: it collects raw usage records from heterogeneous, multi-vendor sources, validates and normalizes them, and delivers structured data to billing, charging, fraud-management, and analytics systems downstream. If that layer drops, duplicates, or mis-transforms records, the damage surfaces as billing errors and revenue leakage. The early Ericsson years were deeply hands-on: C, C++, Java, Linux and Solaris, Oracle and PostgreSQL, multithreading, billing-node integration, and 24×7 production support. By 2019, as Solution Architect and Product Owner for Ericsson Mediation, he had moved into design and automation, building a modular testing framework spanning functional, platform, integration, and regression validation, wired into GitLab CI/CD and run across internal labs, customer-like environments, and Kubernetes targets before code ever reached live traffic. In modern 5G networks the same mediation role is formalized as the Charging Gateway Function, which is why the specialization remains current rather than legacy.
STEEL, RAILS, AND SILICON
After Ericsson, the thread ran through industrial automation: warehouse systems covering ASRS functionality, pallet and container tracking, move management, order processing, fault recovery, and the back-end Equipment Management Software that orchestrates physical machinery, plus a data-load utility that turns messy client inventory files into clean, validated database records. Alongside it came safety-critical rail work in the Positive Train Control domain: test plans, fail-safe scenarios, integration testing, and requirement-to-result traceability for software whose defining duty is to intervene when a human does not.
Today, Verma works on the software that keeps EUV lithography Build/Test systems trustworthy: Python diagnostic and calibration automation, preventive-maintenance scripting, health-check frameworks, root-cause tooling, and regression suites he has refactored deeply enough to cut execution time by roughly twenty percent, per the test logs he insists on citing. Some of the work, such as automation around tin-catch heating control, sits close enough to the hardware that thermodynamics becomes part of the job description.
THE WORK THAT DISAPPEARS
On paper, the project list could read as unusually broad. Read it instead as one continuum. Aircraft test systems taught simulation and disciplined diagnostics. Biometric terminals taught edge reliability and traceability. Satellite broadband taught distributed device management. Mediation taught large-scale data integrity and CI/CD-driven validation. Warehouse automation taught operational software tied to physical movement. Semiconductor equipment brought all of it to bear on some of the most complex machines ever built.
Engineers who build hidden infrastructure rarely produce something a consumer can hold. Their work is measured in quieter outcomes: a test suite that finishes sooner, a health check that catches a thermal anomaly early, a data flow that preserves billing integrity, a utility that prevents an inventory failure. It demands patience (parsing dense logs, reproducing defects that appear only under narrow conditions) and a certain humility, because the best version of the work disappears.
The orientation also points forward. After two decades of reading logs and validating outcomes, Verma has been exploring applied generative-AI tooling, including an in-progress log-file analyzer that combines parsing, retrieval, and summarization to speed root-cause analysis. He presents it modestly, as an extension of an old instinct rather than a reinvention: not replacing engineering judgment, but helping experienced engineers navigate operational evidence faster, with verification intact. Behind every flawless network, every microscopic chip, and every automated facility is an engineer making sure the system earns its trust. That is the career Sudhir Kumar Verma has built.
ABOUT SUDHIR KUMAR VERMA
Sudhir Kumar Verma is a senior software systems engineer and technical lead whose career spans EUV lithography Build/Test automation, Ericsson telecom mediation, warehouse automation, embedded biometric systems, satellite-broadband device management, and defense-grade testing systems. His field is mission-critical software engineering for high-precision semiconductor equipment, industrial automation, and large-scale telecom data platforms. Email: sudhir.veerma@outlook.com | LinkedIn: linkedin.com/in/sv-b866258