Systems Engineer

HARSHITH SUNKU

I don't guess. I trace, profile, and prove it.

WHAT I DO

01

Network Engineering

BGP, routing protocols, packet analysis, and deep inspection. Layer 2 through 7.

02

Kernel & OS Internals

System calls, memory management, scheduler internals, and device driver development.

03

Performance Engineering

Profiling, benchmarking, bottleneck analysis, and systematic optimization.

04

Low-Level Debugging

GDB, strace, perf, eBPF, hardware performance counters, and root cause analysis.

05

Distributed Systems

Consensus algorithms, fault tolerance, distributed tracing, and CAP trade-offs.

06

Systems Security

Network hardening, exploit analysis, secure boot, and defense-in-depth design.

HOW I THINK

I don't trust abstractions I can't explain.

I'm a systems engineer with a deep focus on how computers actually work — not at the framework level, but at the metal. My work spans network engineering, kernel internals, and performance analysis. The most interesting problems live at the boundaries between layers.

I've spent years debugging systems that "shouldn't" be failing — tracing problems from application behavior down through syscalls, kernel code, network stack, and hardware counters. This stack-spanning perspective shapes how I design systems: with explicit attention to every abstraction layer and its failure modes.

I start from first principles. Before reaching for a tool, I understand what the tool does. Before profiling, I form a hypothesis. Before patching, I understand the root cause. This means I'm slower at first and faster over time. Systems I touch tend to stay fixed.

THE PATH

2021 — PRESENT

Senior Systems Engineer

Deep focus on network engineering and kernel-level performance analysis. Root-cause debugging across the full stack from application to hardware.

2018 — 2021

Performance Engineer

Systematic profiling and optimization for high-throughput distributed systems. eBPF, perf, and hardware counter-based analysis.

2015 — 2018

Systems Developer

Network stack development and kernel module work. First encounter with the beauty and horror of concurrent systems at scale.

LET'S CONNECT

Want to talk systems?

harshithsunku@gmail.com
identity.json
{
  "@type": "Person",
  "name": "Harshith Sunku",
  "role": "Systems Engineer",
  "focus": "below the abstraction",
  "approach": "trace → hypothesize → prove",
  "stack": [
    "registers", "syscalls",
    "TCP/IP", "eBPF",
    "...all the way up"
  ]
}