Skip to content
alireza – Home

Open to senior AI / full-stack roles

Alireza RastinehFull Stack AI Engineer

Alireza Rastineh — I turn language models into products that keep working after the demo.

Reliable cloud infrastructure and defensive AI, built to run in production.

See the work

Best for questions about projects, experience, skills and availability.

Answers come from an AI, using this site’s content.

Case studies

Selected work

  • TODO: Five-word descriptor

    TODO: Project One

    TODO: One-sentence hook describing exactly what the product does and who it helps.

    • TypeScript
    • Python
    • FastAPI
    • Postgres
    Details
    Problem

    TODO: Pain point that existed before this software. Quantify the cost (hours, dollars, headcount).

    Architecture

    TODO: Models and AI pipelines used. Name the LLM, the orchestration framework, the retrieval strategy.

    Infrastructure

    TODO: How the AI is served and consumed. Backend framework, database, frontend stack, deployment target.

    Outcomes
    • TODO: Concrete metric #1 (e.g. reduced X by Y%)
    • TODO: Concrete metric #2 (latency, throughput, cost)
    • TODO: Concrete metric #3 (adoption, retention, savings)

  • TODO: Five-word descriptor

    TODO: Project Two

    TODO: One-sentence hook.

    • Angular
    • Node.js
    • Redis
    Details
    Problem

    TODO: Business problem.

    Architecture

    TODO: AI architecture.

    Infrastructure

    TODO: Infrastructure.

    Outcomes
    • TODO: Metric #1
    • TODO: Metric #2
    • TODO: Metric #3

  • TODO: Five-word descriptor

    TODO: Project Three

    TODO: One-sentence hook.

    • React
    • PyTorch
    • Docker
    Details
    Problem

    TODO: Business problem.

    Architecture

    TODO: AI architecture.

    Infrastructure

    TODO: Infrastructure.

    Outcomes
    • TODO: Metric #1
    • TODO: Metric #2
    • TODO: Metric #3

What I work with

Skills

  • frontend + backend + AI

    Core Architecture

    • TypeScript
    • Angular
    • React
    • Next.js
    • Node.js
    • Python
    • FastAPI
    • Hono
    • REST
    • GraphQL
    • WebSockets

    I build responsive interfaces in Angular and React that consume high-throughput Python FastAPI backends, seamlessly routing complex user queries to custom-tuned LLMs.

  • database + API + AI

    AI / ML Stack

    • PyTorch
    • LangChain
    • OpenAI API
    • Anthropic API
    • RAG
    • Vector Search
    • Pinecone
    • pgvector
    • Semantic Caching
    • Prompt Engineering
    • Output Parsing

    I leverage PostgreSQL with pgvector extensions alongside Node.js to create lightning-fast Retrieval-Augmented Generation architectures that deliver accurate, context-aware data to the client.

  • data + ML + cloud

    DevOps & Cloud

    • Docker
    • GitHub Actions
    • Caddy
    • Hetzner
    • Linux
    • Nginx
    • CI/CD

    By orchestrating PyTorch model training pipelines via Docker and GitHub Actions, I ensure that machine learning services deploy cleanly into scalable, highly available cloud environments.

  • storage + streams

    Data Pipelines

    • PostgreSQL
    • Redis
    • Drizzle ORM
    • Prisma
    • Kafka
    • BullMQ

    Typed schemas, vector indexes, and async queues keep inference jobs, embeddings, and user data flowing without bottlenecks.

How I work

About

Building AI applications in a Jupyter notebook is easy; keeping them stable in production is hard. My engineering philosophy centers on defensive AI integration. I treat language models as powerful but unpredictable microservices. By implementing strict semantic routing, output parsing, and automated fallback mechanisms, I actively mitigate hallucinations and enforce reliable data structures. I prioritize the user experience by streaming tokens to the frontend to mask latency, while simultaneously optimizing prompt payloads and leveraging semantic caching to keep API inference costs strictly under control. I don't just write prompts; I engineer resilient, full-stack systems around them.

Alireza Rastineh@alirezarastineh

Based in
Berlin - Remote
Status
Open to senior AI / full-stack roles

Let’s talk

Contact

Email

contact@alirezarastineh.me

Elsewhere

Open to senior AI / full-stack roles

Berlin - Remote

I reply within two working days.

Hiring, or have a project? Check the fit first