@aznikline
systems engineer

The things I once imagined would be my greatest achievements were only the first steps towards a future I can only begin to fathom.

— Jace Beleren

Mason. I work at the boundary of AI agents, quantitative research, and low-level engineering — writing source-first implementations I can read end to end, rather than hiding behind frameworks.

AI agents quant systems OS / kernels OSINT & AI security source-first
selected work
python · ⭐
A source-first study of how agent capabilities accumulate — tool calling → memory → subagents → teams → context compaction → safety, one chapter at a time in ~100 lines of Python each. Read it like a book.
github.com/aznikline/nanoagent →
python · ⭐
A modular Alpha factor mining platform — genetic programming + DeepAlpha + a WorldQuant-style expression engine — across A-share and US markets. CLI, Streamlit, and notebook-first APIs.
github.com/aznikline/alpha-mining-system →
python · ⭐
An engineered interview library, not a link dump — original handbook content + training tracks, backed by a toolchain (content index, doc linter, progress dashboard) and passing tests.
github.com/aznikline/interview →
An async Rust client for Binance Spot REST APIs, inspired by python-binance — focused on the most valuable Spot surface first rather than chasing full parity.
github.com/aznikline/rust-binance →
the thread
now
AI safety & OSINT tooling. Contributing to maigret (33k-star OSINT dossier tool) — 2 PRs merged, maintainer called the work “amazing” and “very precise”. Also following shannon (white-box AI pentester).
recent
AI agents. nanoagent — a compact reference for what an agent actually needs to act, remember, delegate, coordinate, compress, and stay safe.
earlier
Quantitative research. Alpha factor mining systems and backtesting pipelines — genetic programming and deep factor discovery.
roots
Systems & vision. An aarch64 microkernel (capability-based IPC, scheduler, memory) and early academic work on visual object tracking.
elsewhere