Zohreh Sadeghi

Zohreh Sadeghi

Fullstack Developer — Malmö, Sweden

Job Fit Analyzer

Paste a job description and get a grounded fit report — generated from my actual project, skill, and experience data via retrieval-augmented generation, not a generic AI wrapper.

0 / 8000 characters (min 50)

Projects

AI-Integrated Creative Assistant

A streaming AI agent chat feature built with the Vercel AI SDK, pairing server-side streamText with client-side useChat for token-level, real-time response rendering. Implemented five server-side tool definitions (board search, pin retrieval, colour-palette lookup, inspiration summaries) letting the model autonomously chain up to five tool-calling steps per response, integrated directly into a production-style app.

  • –Five chainable server-side tool definitions for agentic, multi-step responses
  • –"Save to Boards" action routes AI-generated suggestions directly into the app's creation flow
  • –localStorage-backed conversation persistence, synced on stream finish
  • Next.js
  • TypeScript
  • Vercel AI SDK
  • OpenAI API
  • React

AI-Powered Developer Portfolio (this site)

Code

A fullstack portfolio built as a real production system rather than a template: a Next.js/TypeScript frontend on Vercel, a standalone NestJS API on AWS ECS Fargate, and a Postgres database managed with Prisma. The flagship feature is a Job Fit Analyzer — visitors paste a job description and get a grounded fit report generated via retrieval-augmented generation (RAG) over the site's own project/skill/experience data, with citations, a match score, and gap analysis, rather than a generic chatbot wrapper. Includes a self-built MCP server exposing project/skill/experience data as callable tools, and is designed security-first given it accepts untrusted text into an LLM pipeline: input validation, rate limiting, and prompt-injection hardening protect the endpoint.

  • –Job Fit Analyzer: RAG-grounded fit report generation, not a generic chatbot
  • –Self-built MCP server exposing personal project data as agent-callable tools
  • –Dockerized NestJS API deployed to AWS ECS Fargate via Terraform-managed infra
  • –Security-first design: zod validation, rate limiting, and prompt-injection hardening around the LLM endpoint
  • Next.js
  • TypeScript
  • NestJS
  • Prisma
  • PostgreSQL
  • pgvector
  • Docker
  • AWS (ECS Fargate, RDS, Secrets Manager)
  • Terraform
  • GitHub Actions
  • Vercel AI SDK
  • OpenAI API
  • MCP

Experience

Fullstack Developer — Insighta Inc.

Mar 2025 – Present

Delivered production features end-to-end on a two-week agile release cycle while the company doubled its client base. Built and scaled a React/TypeScript frontend (interactive dashboards, sortable data tables, dynamic filters) turning complex datasets into self-serve tools for non-technical business users. Partnered cross-functionally with product managers, UX designers, and backend engineers, shaping API design. Owned CI/CD pipelines and deployment workflows on AWS; raised engineering standards through code reviews. Stack: TypeScript, React, Next.js, Redux, Tailwind CSS, Node.js, Jest, Playwright, REST APIs, GitHub Actions, AWS.

System Development — Security Focus (Higher Vocational Education) — Jensen Yrkeshögskola

Aug 2024 – Jun 2026

Full stack application development in TypeScript, React, C#, .NET, Java, and SQLite. REST API design and implementation; client-side and server-side architecture for scalable applications.

Fullstack Developer — Sigma Technology Embedded Networks

Dec 2023 – Mar 2025

Built and shipped user interfaces for client applications on the Axis Camera Application Platform (ACAP), ramping up quickly in a specialised embedded systems domain. Delivered real-time dashboards and traffic-analysis tooling from live IoT device data, working directly with backend and firmware engineers on API integration and data flows. Managed CI/CD pipelines and refactored core code paths, improving performance and long-term maintainability.

Frontend Developer (Intern) — Polestar

Dec 2022 – Dec 2023

Delivered client-facing web application features in an agile, cross-functional team, with a focus on quality, usability, and clean code. Applied TDD principles and contributed to CI/CD pipeline maintenance across sprints.

DevOps Engineer Programme (Higher Vocational Education) — Jensen Yrkeshögskola

Aug 2022 – Jun 2024

Cloud system design on AWS and Azure; Kubernetes, Docker, CI/CD pipelines, Terraform and GitHub Actions. Infrastructure automation, monitoring, and observability in cloud-native environments.

Skills

AI & LLM Engineering

  • Vercel AI SDK
  • Streaming AI UIs
  • Agentic Tool-Calling
  • OpenAI API
  • Prompt Engineering
  • RAG (Retrieval-Augmented Generation)
  • Model Context Protocol (MCP)
  • Prompt-Injection Hardening

Backend

  • Node.js
  • REST API Design
  • ASP.NET / .NET
  • NestJS
  • Prisma ORM

Cloud & DevOps

  • Docker
  • AWS
  • GitHub Actions / CI-CD
  • Docker Compose
  • Kubernetes
  • GCP
  • AWS Secrets Manager
  • Terraform (IaC)
  • AWS ECS Fargate
  • AWS RDS

Data Visualisation

  • Chart.js
  • Recharts
  • D3.js

Databases

  • SQLite
  • PostgreSQL
  • pgvector

Design & Collaboration

  • Code Reviews
  • Agile/Scrum
  • Figma

Frontend

  • Next.js
  • React
  • Responsive Design
  • Component Architecture
  • Accessibility (WCAG)

Languages

  • TypeScript
  • JavaScript (ES6+)
  • HTML5
  • CSS3
  • Java
  • C#

Security

  • Input Validation (Zod)
  • Rate Limiting
  • Least-Privilege IAM

State Management

  • Redux
  • React Context

Testing & QA

  • Jest
  • ESLint
  • Playwright
  • React Testing Library
  • Cypress