/* profile.kartik-sirohi */

Kartik Sirohi

Performant AI systems — IT undergrad at NIT Jalandhar
Kartik Sirohi
kartik.orion.dev@gmail.com github linkedin twitter
@@ publications @@
P²INNs system diagram
P²INNs: Physio-Psychiatry Informed Neural NetworksIEEE Access, 2026
A novel architecture for depression risk prediction integrating physiological and psychiatric indicators. Developed during a research internship at the University of Galway.
ieee xploredoi
@@ experience @@
f3a9c21 Aug 2026 – present HEAD
AI Engineering Intern @ Horrazon AI
b7e401d Jul – Aug 2026
Research Intern @ QWorld
1c88ae4 Jan – Mar 2026
AI Research Intern @ University of Galway
@@ open source @@
KS llama.cpp 2 PyTorch 1 OpenSRE 3 alexandria3k 1 SciPy 1
hover or focus a node — 8 merged patches across 5 repositories
llama.cpp merged concat row-level memcpy — #24575 SIMD vector kernel — #22209
PyTorch merged NaN clamp fix — #172200
OpenSRE merged unit tests & validation — #852 #917 Tempo service tests — #882
alexandria3k merged O(n)→O(1) lookup — #63
SciPy merged Gauss-Jacobi docs — #23129
@@ projects @@index / 01–10
systems
tinyGPT mini-torch linalg cuda kernels
agents
Rocky Athena maze-runner
experiments — from Fun
minimax http-server FileManager
hover or focus a tile for details
@@ posts @@
From 44 GFLOP/s to 376 GFLOP/s: Learning GPU Performance Through MatMulSep 2026 · Research
Notes on writing and tuning CUDA matmul kernels.
Building tinyGPT in C++Jul 2026 · Projects
Implementing a 30M-parameter GPT from scratch with NEON, SSE, and OpenMP.
Backpropagation: The Engine of Deep LearningJan 2026 · Tutorials
An intuitive walkthrough of backpropagation — from chain rule to automatic differentiation.
Intuition Behind Neural NetworksOct 2025 · Tutorials
Understanding neural networks one parameter at a time.
Graph Retrieval MethodsJul 2026 · Research
A survey of why graph retrieval methods are useful for retrieving knowledge.