Skip to content
View oneKn8's full-sized avatar
🎯
Focusing
🎯
Focusing
3 followers · 1 following

Achievements

Achievement: Pull Sharkx2Achievement: QuickdrawAchievement: YOLOAchievement: Pair Extraordinaire

Achievements

Achievement: Pull Sharkx2Achievement: QuickdrawAchievement: YOLOAchievement: Pair Extraordinaire

Highlights

  • Pro

Block or report oneKn8

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
oneKn8/README.md

Santo

I build the infrastructure most engineers import.

Vector search engines, distributed task queues, voice AI pipelines, quantum circuit optimizers -- from scratch, in C++, Go, Rust, Python, and TypeScript.

CS @ UT Dallas | CTO @ Rhemic -- AI-powered search visibility platform

Python TypeScript C++ Go Rust Java

VectorVault

Vector search engine from scratch in C++20. Sub-millisecond queries at 100k vectors. No FAISS. No external libraries.

HNSW approximate nearest neighbor search with AVX2 SIMD-accelerated distance computation and memory-mapped persistence. Built to understand how vector databases actually work under the hood.

C++20 CMake Docker

Soniq

AI voice agents that pick up business phones, book appointments, and handle objections -- all in under 500ms.

Full-stack voice pipeline: Deepgram STT, LLM orchestration, Cartesia TTS. Multi-tenant SaaS with CRM integration, calendar sync (Google, Outlook, Calendly), and 27 industry-specific agent presets.

TypeScript Next.js Python LiveKit Supabase

QUASAR

What if an LLM could discover quantum circuits?

Physics-augmented LLM proposes candidate circuits. A surrogate neural network filters 10,000+ candidates in milliseconds. The top 10% go through full VQE quantum simulation. Bridges machine learning with quantum chemistry.

Python PyTorch Qiskit

More Work

Project What Built With
RivetQ Durable task queue with WAL, Raft consensus, dead-letter queues, rate limiting Go, gRPC, Prometheus
Research Agent Autonomous research pipeline with locally-hosted DeepSeek-R1-14B via vLLM Python, LangGraph, FastAPI
OpenLock Open-source lockdown browser for Linux -- SEB protocol, kiosk shell, VM detection C++17, Qt6, Chromium
Pulse HPC cluster observability -- GPU telemetry, SLURM scheduling, LLM ops assistant Go, Prometheus, Grafana
SearchLight Hybrid search API (BM25 + HNSW) via Lucene 9, hexagonal architecture Java 21, Spring Boot
Socratic Private multi-model NotebookLM alternative -- 16+ providers, vector search, podcasts TypeScript, Next.js
Syzygy Document engineering IDE -- Rust + Tauri 2.0 + Typst backend, visual pipelines Rust, Tauri, Typst
mcp-memory Persistent agent memory server with semantic search and importance scoring Python, ChromaDB, MCP

Contribution Graph

X LinkedIn Email Discord GitHub

Pinned Loading

  1. RivetQ RivetQ Public

    Durable task queue with write-ahead log, Raft consensus for HA, priority queues, dead-letter queues, rate limiting. REST + gRPC APIs, Prometheus metrics, Next.js admin UI.

    Go 1

  2. VectorVault VectorVault Public

    HNSW approximate nearest neighbor engine from scratch in C++20. AVX2 SIMD-accelerated distance computation, memory-mapped persistence, sub-millisecond queries at 100k vectors.

    C++ 1

  3. openlock_browser openlock_browser Public

    Open-source lockdown exam browser for Linux. SEB protocol implementation (BEK + Config Key + RNCryptor v3), kiosk shell, process guard with cgroups, VM/debugger detection.

    C++ 5 1

  4. searchLight searchLight Public

    Hybrid search API combining BM25 keyword matching and HNSW vector similarity via Lucene 9. Hexagonal architecture, Spring Boot 3.3, Prometheus + OpenTelemetry.

    Java 1

  5. socratic socratic Public

    A private, multi-model, full-featured alternative to Notebook LM

    TypeScript 1

  6. soniq soniq Public

    AI voice agents for business phone handling. Sub-500ms latency, multi-tenant, 27 industry presets.

    TypeScript 1