Agentic AI/ML · Quantitative Finance

Systems for models, markets, and scale.

I'm Zhanyl Abdybaeva, an ML infrastructure and quantitative finance engineer working at a quantitative trading firm. My work spans GPU clusters, inference systems, distributed training, and the schedulers that decide what runs when.

This site is where I publish long-form technical writing, shorter field notes, and project pages for the systems I'm building.

GitHub LinkedIn Résumé About RSS

Role

Infrastructure engineering at a quantitative trading firm

Focus

GPU clusters · inference · Slurm at scale

Studying

MS CS — ML @ Georgia Tech
CQF

Writing

Deep dives & lab notes
most weeks

New writing most weeks · last published July 26, 2026

ls ./writing --featured

January 2026 · blog · 8 min read

LServe and SampleAttention: What Sparse Attention Actually Changes in Prefill and Decode

Structured sparsity aligned with GPU attention kernels: LServe’s unified serving stack for prefill and decode, and SampleAttention’s empirical patterns plus CRA as a runtime quality floor.

January 2026 · blog · 4 min read

How KV-Cache Paging Works in vLLM — and Why It Matters for Production

A technical walkthrough of PagedAttention: the memory management innovation that makes vLLM practical for serving many concurrent LLM sessions without fragmenting GPU memory.

All deep dives →

tail ./lab-notes

All lab notes →

ls ./projects

Epilog GPU Validator Shipped
Drain a node for a persistently faulty GPU — and never for a transient one.
slurmgpunvidiahpc
Source →
Agentic ML Inference Infrastructure In Progress
Production agentic AI infrastructure layer: autonomous systems that reason, decide, and act on real workloads using vLLM, LangGraph, and MCP.
inferencevllmagentic-ailangraph
Slinky GitOps Shipped
Slurm on Kubernetes from nothing in one command — and the auth-key rotation nobody wants to test in production.
slurmkubernetesslinkyhelm
Source →
128-GPU Distributed Training Platform In Progress
Distributed ML training infrastructure for large-scale models: FSDP, gradient checkpointing, mixed precision, and cost-optimized multi-node orchestration.
distributed-trainingfsdppytorchgpu
IB Slurm Exporter Shipped
Correlate InfiniBand and RoCE counters with the Slurm job that owns them, so a slow collective can be traced to the fabric.
slurminfinibandrdmaprometheus
Source →
GPU-Accelerated Volatility Surface Calibration Design Phase
CUDA-accelerated local and stochastic volatility surface calibration — Dupire's model, Heston, and hybrid approaches for real-time derivatives pricing.
quantitative-financecudagpuvolatility
GPU Reaper Shipped
Find wasted GPU allocations on a Slurm cluster and escalate them through alert, drain, and cancel.
slurmhpcgpunvidia
Source →
Slurm Scheduler Lab Shipped
Test Slurm priority and backfill policy against a job trace before it reaches a live controller.
slurmhpcschedulingbackfill
Source →