Skip to content
View ns-0437's full-sized avatar
🙃
🙃
5 followers · 7 following

Achievements

Achievement: YOLOAchievement: Pull Shark

Achievements

Achievement: YOLOAchievement: Pull Shark

Block or report ns-0437

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
ns-0437/README.md
  • 👋 Hi, I'm @ns-0437 (Navin Kumar)
  • 👀 I'm interested in retrieval systems: chunking strategies, hybrid dense + BM25 fusion, HyDE, cross-encoder reranking, metadata channels, and the evaluation that decides which of them earn their place
  • 🤖 I also work on production ML and the machinery that keeps it honest: EmbeddingGemma-300m, ChromaDB, RT-DETR, PyTorch, FastAPI, Docker, Cloud Run
  • 🌱 I work as an AI Engineer at DigitalXcode, shipping agentic pipelines that run unattended
  • 💞️ I'm looking to collaborate on open source around agent reliability, supervision and observability for long-horizon systems
  • 📫 How to reach me navinkumar0437 at gmail

Pinned Loading

  1. agentfuse agentfuse Public

    A logical circuit breaker for long-range autonomous agents: detects tool loops, goal drift, logic traps and runaway spend, then steers the agent back with a separate reasoning model. Works with Ope…

    Python 2

  2. relevant-section-identification relevant-section-identification Public

    Given a PDF and a natural-language query, identify the pages and section headings that answer it. Hybrid retrieval over structure-aware chunks, with a PDF-viewer UI.

    Python 2

  3. research-scout research-scout Public

    Post-meeting research agent for the SitRep marketplace: mines meetings for open questions, competitors, and claims, then runs live web research and returns a sourced briefing.

    Python 1

  4. Human-Activity-Recognition-in-Dark Human-Activity-Recognition-in-Dark Public

    Human activity recognition in low-light video using an LRCN (CNN + LSTM) model on the ARID dataset

    Jupyter Notebook 2