Skip to content
View schart's full-sized avatar
πŸ₯·
Wizard
πŸ₯·
Wizard

Block or report schart

Report abuse

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

Report abuse
schart/README.md

Heja Arslan

AI Infrastructure Engineer Β· Turkey

I build the infrastructure that makes AI systems run in production β€”
pipelines, vector search, event-driven ingestion, and scalable backends that don't break under load.

LinkedIn   Email


πŸ‘¨β€πŸ’» About Me

AI Infrastructure engineer focused on building the systems that sit beneath AI products β€” the data pipelines, vector stores, async workers, and query engines that make intelligence scalable and reliable.

  • 🧠 Built a custom DSL (query language) with parser & lexer from scratch β€” used in production
  • πŸ“‘ Event-driven pipelines: RabbitMQ, async ingestion, GraphQL subscriptions
  • πŸ” Vector search & AI retrieval: Pinecone, embedding pipelines, semantic matching
  • ☁️ Deployed on AWS (CloudFront), containerized with Docker, time-series with InfluxDB
  • πŸ”§ Core stack: TypeScript / Node.js / NestJS β€” ships fast, stays clean
  • 🌍 Open source contributor β€” vscode-material-icon-theme

πŸ›  Tech Stack

Languages

TypeScript Python

Backend & Frameworks

NestJS Node.js GraphQL RabbitMQ

AI Infrastructure

Pinecone InfluxDB Sequelize Prisma

Infrastructure

Docker AWS CloudFront Stripe


🌍 Open Source

Project Role
FlexQL Author & Maintainer ⭐ 30
vscode-material-icon-theme Contributor ⭐ 2.7k

πŸ“Œ Projects

FlexQL ⭐ 30

Custom query language (DSL) for flexible, secure data filtering β€” built without heavy SQL or ORM abstractions.

Full parser & lexer from scratch. Expressive syntax, safe execution, easy integration into any Node.js backend. Demonstrates deep understanding of language design and compiler fundamentals.

Stack: TypeScript Β· Custom Parser/Lexer


AI-powered streaming aggregator β€” surfaces free and cheapest available sources for movies & series.

Async content ingestion via RabbitMQ, edge delivery via CloudFront, semantic source matching via Pinecone vector search, time-series analytics in InfluxDB, Stripe monetization. Query engine powered by FlexQL.

Stack: NestJS Β· RabbitMQ Β· AWS CloudFront Β· InfluxDB Β· Pinecone Β· Stripe


Full-stack notification service β€” built and shipped end-to-end in 4 days.

GraphQL API, RabbitMQ event bus, Prisma ORM, Docker deployment.

Stack: GraphQL Β· RabbitMQ Β· Prisma Β· Docker


Jira-inspired project management tool with kanban workflow (Todo β†’ In Progress β†’ Review β†’ Done).

Stack: NestJS Β· TypeScript


Real-time uptime monitoring service β€” tracks service health and liveness over time.


πŸ“Š GitHub Stats

 


Open to AI infrastructure

Pinned Loading

  1. FlexQL FlexQL Public

    FlexQL – A powerful, secure, and extensible query language for flexible data filtering without complex SQL or ORM queries.

    TypeScript 30 1

  2. NotiFlow NotiFlow Public

    This fully functional notification service was designed and implemented in just 4 days, demonstrating rapid prototyping and full-stack skills with GraphQL, RabbitMQ, Prisma, and Docker

    TypeScript 2

  3. uptime-livenliness uptime-livenliness Public

    Real-Time uptime data monitoring

    TypeScript 3

  4. TaskSphere TaskSphere Public

    TaskSphere an application similar like Atlassian/Jira, breaks project into small pieces to make management easy (Todo, InProgress, Done, Review).

    TypeScript 3