About
Systems programmer who keeps asking "but why is it slow?" until the answer is in CPU cycles.
I'm a systems programmer based in India, currently at Qualys working on macOS agent security and Software Composition Analysis. Before that, I interned at SentinelOne reverse-engineering macOS malware and writing detection rules — which is where I learned that understanding the attacker means understanding the machine.
But the thing that keeps me up at night is performance. Not the "make it faster" kind — the "understand exactly why the CPU is stalling on this instruction" kind. x86-64 microarchitecture, cache hierarchies, branch prediction, the memory bus. I built PicoPerf because I wanted to stop guessing and start measuring — retired instructions, cache misses, branch mispredictions, all as independent hardware signals at single-cycle resolution.
Through Low-Latency-Lab,
I take my old projects apart and rebuild them with everything I now know about the hardware. io_uring async I/O,
AVX2 SIMD, lock-free queues, O_DIRECT, CPU pinning, PGO — then profile with perf and optimize until
the silicon is the bottleneck, not my code. The fun part? Finding a kernel thread silently stealing 40% of a
core, or discovering that a single PAUSE instruction can cut 48% of wasted instructions. Those
moments are better than coffee.
When I'm not writing assembly or staring at perf stat output, I'm reading about physics,
mathematics, or philosophy. Usually at 3am. The obsession is real.
Skills
Languages
- C,
- C++,
- x86-64 Assembly,
- CUDA C
Low-Latency & Performance
- SIMD/AVX2 Intrinsics,
- Branchless Programming,
- Cache-Friendly Programming,
- LTO/PGO,
- Pipeline Optimization,
- Lock-free Data Structures,
- Async I/O (io_uring),
- O_DIRECT & Zero-copy I/O
CPU and x86 specific skills
- PMU Hardware Counters,
- ILP & Out-of-Order,
- Cache Hierarchy,
- TLB Internals,
- RDTSC/RDTSCP
Systems & Tools
- Linux perf_event API,
- GDB,
- Valgrind/Callgrind,
- perf stat/record/annotate,
- io_uring,
- Shared Memory (mmap/shm),
- macOS Security
Experience
Software Developer — Qualys
MacOS agent security, CVE patching, SCA integration, custom macOS scanners, CPU usage limiter for agent performance.
Security Software Intern — Sentinel One
Reverse engineered macOS malware, authored 4 new detection rules, optimized rule logic achieving up to 86% false positive reduction.
B.Tech Computer Science — PES University
Best Final Year Project 2024 for post-quantum cryptography research. Winner of mathematics hackathon, top 5 in 3 CTF events.
Contact
Open to low-latency developer, systems, and HPC roles.