About Growtrics Growtrics is an AI-powered EdTech company building an adaptive learning platform that helps students master Mathematics and Science through personalised, interactive, and deeply engaging experiences. We combine cutting-edge AI models, multimodal learning, and delightful product design to deliver a new category of learning outcomes.
We are expanding our backend and AI infrastructure capabilities, and are looking for an exceptional Backend Engineer who can build production-grade OCR + LLM multi-agent systems to power our automated question extraction and grading pipelines.
Role Overview
As a Backend Engineer at Growtrics, you will design and implement a multi-agent Python pipeline that converts unstructured exam papers (PDFs) into structured question-answer schemas consumable by our student app. You will work with FastAPI, LLMs, OCR/Vision models, LangChain-like agent frameworks, and scalable cloud infrastructure.
This role sits at the intersection of GenAI engineering, backend systems design, and EdTech product development.
Key Responsibilities:
Backend Engineering (FastAPI & Pipeline Architecture);
- Design and implement a modular, multi-stage Python pipeline for OCR, content extraction, parsing, and structuring exam questions.
- Own and productionise FastAPI microservices that interact with LLMs, OCR engines, and internal data stores.
- Build robust logging, monitoring, and cost-tracking systems for all agents/stages.
GenAI, LLMs, Multi-Agent Systems:
- Build multi-agent orchestration pipelines (e.g., LangChain, LlamaIndex, or custom frameworks).
- Implement LLM-based content structuring, extraction, and reasoning workflows.
- Model and manage agent input/output, reasoning traces, and message schemas.
- Ensure outputs are consistent, structured, and generalizable across diverse exam formats.
OCR + Computer Vision:
- Integrate OCR / Vision models for extraction of text and embedded images from PDFs.
- Handle multimodal processing (text, diagrams, equations, tables).
- Ensure reliable extraction of mathematical notation, diagrams, graph-based questions, etc.
Question Schema & Data Modelling:
- Design scalable schemas representing:
- Questions, sub-questions (parts a/b/c), and answer keys
- Images, diagrams, and mathematical expressions
- Grading metadata supporting LLM-based evaluation
- Ensure output schemas serve both UI rendering & automated grading. 5. Quality, Observability & Reliability
- Implement structured logging for each agent step using datetime-tagged logs.
- Build the monitoring layer to track failure rates, false positives, and pipeline health.
- Implement automated testing for multi-stage pipelines.
Technical Requirements:
Core Skills:
- Strong proficiency in Python 3.9+
- Experience building production microservices using FastAPI
- Deep hands-on experience with LLMs, prompt engineering, and agentic workflows
- Understanding of OCR engines (e.g., Tesseract, PaddleOCR, Vision-Language models like GPT 5, Gemini 3 Pro, Qwen3VL etc.)
- Experience with PDF parsing (e.g., pymupdf, pdfplumber, unstructured)
- Familiarity with LangChain BaseMessage schemas or equivalent custom agent frameworks
- Ability to structure multi-stage pipelines with clean code organisation.
Software Engineering Best Practices:
- Strong in modular design, documentation, and maintainability
- Experience implementing robust logs, monitoring, and analytics
- Familiarity with cloud environments for scaling inference workload
- Experience with CI/CD, unit tests, and integration tests
- Experience working with structured/unstructured data and data modelling
Bonus Skills:
- GenAI vector DBs (Pinecone, Weaviate, Chroma)
- Multi-agent orchestration tools
- Educational content processing or assessment system
- Knowledge of math or science content is a plus (not required) What You Will Build (Example Projects) You will directly contribute to:
- Automated OCR Question Schema Answer Key pipelines
- LLM-powered grading systems for handwritten or digital responses
- Multi-agent systems with reasoning trace logging
- Scalability frameworks that enable Growtrics to ingest 10,000+ exam papers across syllabi and countries
- Multimodal processing for diagrams, graphs, and equations
Why you'll love working here:
- Highly competitive compensation package, including 100% salary during probation, designed to reward your talent and dedication.
- 13th-month salary and performance-based bonuses to celebrate your achievements and contributions.
- Comprehensive Social Insurance calculated on gross salary, giving you peace of mind and full legal protection.
- Generous and flexible time-off policy: 14 annual leave days + 6 sick leave days to recharge and maintain work-life balance.
- Flexible working hours allowing you to manage your schedule and work in a way that suits your lifestyle.
- MacBook provided, ensuring you have the tools to perform at your best from day one.
- Full compliance with the Vietnam Labor Code, so you can focus on work with confidence in a fair and lawful environment.
- Opportunity to collaborate directly with global stakeholders, gaining exposure to international best practices and expanding your professional network.
- A workplace that truly values your well-being: regular social events, sports clubs, gym activities, and team-building programs to foster connection and fun.
- Vibrant, youthful, and international culture that encourages creativity, innovation, and continuous growth.
- Fully stocked pantry with a variety of snacks, milk, and beverages to keep you energized throughout the day.
- Recognition and rewards based on your experience, skills, and qualifications, ensuring your contributions are appreciated and fairly compensated.
- Career development and learning opportunities to help you grow personally and professionally in a dynamic, supportive environment.
