Skip to content
View macayu17's full-sized avatar
:shipit:
:shipit:

Highlights

  • Pro

Block or report macayu17

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 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
macayu17/README.md

Banner

Typing SVG

Ayush Kumar

I build backend, real-time, and AI-integrated systems across distributed simulation, AI memory infrastructure, trading workflows, and open-source tooling.

Portfolio LinkedIn Twitter Email


Open Source Activity

I am building my open-source track across developer tooling, AI infrastructure, data systems, documentation, and creative coding projects.

Area Repositories and contribution targets
Developer tooling Vitest, pyenv, OpenTelemetry JS
AI and agents Pydantic AI
Web and frameworks WordPress, Svelte
Data and infrastructure CDLI on GitLab, NASA F Prime GDS, OpenTelemetry

Current snapshot

  • 76 public open-source contributions across GitHub and GitLab (64 GitHub PRs and 12 GitLab MRs), with 39 merged.
  • Contributed 12 merge requests to CDLI on GitLab, with 9 merged, improving repository tooling and data-processing workflows.
  • Opening PRs across OpenTelemetry JS, Vitest, Svelte, WordPress, Pydantic AI, pyenv, NASA F Prime GDS, and other tooling repos.

Featured Projects

Project Focus Stack
Engram Self-hostable AI memory layer that retrieves, injects, extracts, and syncs durable user context across LLM clients. Live demo FastAPI, pgvector, MCP, Next.js, TypeScript
SENTINEL Multi-agent market microstructure simulator for liquidity-crisis prediction, large-order detection, and live order-book analytics. Live demo Python, FastAPI, Next.js, WebSockets, XGBoost, scikit-learn
EquityFlow Real-time paper-trading simulator for stocks, F&O, and commodities with broker data, portfolio tracking, and trading-desk controls. Live demo Next.js, React, TypeScript, FastAPI, Python, broker APIs
Parkinson's Disease Screening Transformer-based clinical classification pipeline trained on 42k+ patient records with a Flask inference API. PyTorch, Transformers, LightGBM, Flask, CUDA
Occasio Event-management and booking platform with publishing, payments, QR tickets, check-in, analytics, teams, and certificates. Live demo React, Vite, Node.js, Express, Prisma, PostgreSQL, Redis

Tech Stack

Languages

Python TypeScript JavaScript C++ Java C

Backend, APIs, and Frontend

FastAPI Node.js Express REST APIs WebSockets Next.js React Tailwind CSS

AI / ML

PyTorch Transformers scikit-learn LangChain OpenAI API NumPy Pandas

Databases, Cloud, and Tools

PostgreSQL MongoDB AWS Azure Docker


GitHub Dashboard

Ayush Kumar's GitHub stats Ayush Kumar's top languages

Ayush Kumar's GitHub streak

Ayush Kumar's contribution graph


Currently Exploring

  • AI memory systems, LLM tool calling, MCP clients, and RAG pipelines.
  • Real-time trading systems, market microstructure, and low-latency dashboards.
  • Distributed systems, event-driven architecture, queues, and production-style observability.
  • Open-source CLI tooling, developer experience, and infrastructure workflows.

Profile views

Pinned Loading

  1. Engram Engram Public

    Engram is a self-hostable AI memory layer. It sits between an application and an LLM provider, retrieves relevant user memories from pgvector, injects them into chat prompts, forwards the request, …

    Python

  2. SENTINEL SENTINEL Public

    Sentinel is a real-time market microstructure simulator built to study liquidity shocks, large hidden orders, and institutional activity through a multi-agent trading environment.

    Python 3

  3. Equityflow Equityflow Public

    EquityFlow is a real-time paper trading platform for stocks, derivatives, and commodities, designed around live market data, portfolio accounting, and execution flows that mirror trading behavior w…

    TypeScript

  4. PRISM PRISM Public

    This project focuses on text-based Parkinson's disease screening using transformer models and ensemble learning on roughly 42,000 patient records, packaged as a clinical assessment portal.

    Python