2024 · Chennai · Professional certification
Full time on the systems stack from the bottom up - C and data structures, Linux internals, ARM and 8051, and embedded C on real silicon. This is where socket programming and RTOS scheduling stopped being coursework and became things I had actually shipped.
Vector India is a professional training institute running embedded systems and full-stack programmes with a dedicated placement cell. It operates from Ameerpet in Hyderabad, Jayanagar in Bangalore and Perungudi in Chennai, plus an online track. I took the Chennai programme.
The reason it mattered for me is the shape of it: full time, on hardware, with an assessed project at the end. It is closer to an apprenticeship than a course, and the parts of my current work that involve Linux internals and concurrency trace directly back to it. Vector India advertises the programme as a six-month course; my own certificate records May to December 2024, which is the period I actually attended.
Programme structure and institute details on this page are from Vector India's own course page, vectorindia.org. The dates and the capstone are mine.
My certificate names the coverage as C & DS, Linux Internals, ARM, 8051 and Embedded C. The institute publishes a wider syllabus for the programme, and the TCP/IP and real-time Linux material below is part of it - taught in order, each module building on the one before.
Pointers, dynamic memory and file handling; data structures, sorting and searching; then C++ - object orientation, inheritance, polymorphism and exception handling. The memory discipline here is the same discipline the arena allocator work later needed.
Kernel architecture and system calls, process management, interrupts and signals, file management, virtual memory, and the full IPC set - pipes, FIFOs, message queues, shared memory and semaphores. Plus multithreading and shell scripting.
The OSI and TCP/IP layers, routing, and the application protocols on top - FTP, TELNET, DNS, DHCP, SMTP. Then socket programming and concurrent servers, which is the direct ancestor of the thread-pooled TCP server I wrote afterwards.
RTOS fundamentals, kernel compilation, applying the RT-PATCH, the real-time API, latency measurement and porting to ARM. Measuring latency rather than assuming it is a habit that started here.
Intel 8051 architecture and embedded C, on-chip peripherals - ports, timers, UART. And the I2C and SPI protocols. Then the ARM7 instruction set on an LPC2114 under Keil.
The programme ends in a build. Mine was a multi-node body control module over CAN - bus arbitration, message scheduling and fault handling under RTOS timing constraints.
Issued by Vector India on 19 March 2025, naming the course, the coverage and the period attended. The wording below is theirs, quoted as it appears.
Three things from this programme are load-bearing in what I do now. Socket programming and concurrent servers became the multithreaded TCP server. Linux internals - specifically IPC and virtual memory - became the reason I could reason about the memory core in Cerberus. And the CAN capstone is still the clearest example I have of a distributed system small enough to hold in your head completely.