Skip to content
View danieljhkim's full-sized avatar
🎯
Focusing
🎯
Focusing

Organizations

@constellation-works

Block or report danieljhkim

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.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

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

Report abuse
danieljhkim/README.md

Daniel Kim

Backend and infrastructure engineer in San Jose. By day: distributed systems, search, and data infrastructure. After hours: tooling that lets AI coding agents do real work safely.

Now building constellation-works, a small set of local-first tools for working with coding agents.

  • orbit: a delivery layer for AI coding agents. An agent files work as a task; Orbit runs it in a sandboxed worktree with file locks and review gates, and every run ends in a pull request with an audit log. Works with Claude Code, Codex, Cursor, Copilot and others. brew install constellation-works/tap/orbit
  • pulsar: an MCP server that lets agents post to X as a human-authorized account without ever holding the credentials.
  • Work in progress: nebula (idea lineage graph), orbit-graph (code graph indexing), orbit-research (reproducible research workflows).

Systems projects

dsearch Distributed search engine: BM25, vector, and hybrid ranking over sharded Lucene indices
kvDB Distributed key-value store in Java with shard routing, replication, and a control plane
local-data-platform Local HDFS/YARN + Hive + Spark dev environment with profile-based config overlays
monodev Reusable dev overlays (agent instructions, scripts, editor config) without polluting git

Background: distributed systems and fault tolerance · search infrastructure (Lucene, embeddings, ranking) · Java and Go backend services · data platforms (Spark, Airflow, Hive, Hadoop, GCP, AWS).

danieljhk.com

Pinned Loading

  1. constellation-works/orbit constellation-works/orbit Public

    Local-first delivery layer for AI coding agents. Your agent files the work as a task, and Orbit runs it in a sandboxed worktree with file locks and review gates. Every run ends in a pull request, w…

    Rust 11 3

  2. data-structures-algorithms data-structures-algorithms Public

    Data Structures and Algorithms

    Python 1

  3. kvDB kvDB Public

    A distributed key-value database with shard routing, replication, and a dedicated control plane, built in Java.

    Java

  4. dsearch dsearch Public

    A distributed search engine supporting BM25, vector search, and hybrid ranking over sharded Lucene indices.

    Java 2

  5. monodev monodev Public

    CLI for managing reusable dev overlays (agent instructions, dev scripts, editor config) across different sessions and branches without polluting git.

    Go 5

  6. local-data-platform local-data-platform Public

    Local Hadoop (HDFS/YARN) + Hive + Spark dev environment manager with profile-based config overlays.

    Go 5