ROLE OVERVIEW

We are looking for an AI Engineer who is excited to build real, impactful AI systems — not just call APIs. You will be the AI layer powering TeenCare‘s core product: translating what happens in weekly teen-mentor sessions into structured behavioral intelligence that parents can actually act on.

You will work side-by-side with our Full-Stack Engineer to embed AI deeply into the product — from the models and pipelines you design to the features families experience every day.

If you love experimenting with LLMs, care about clean pipelines, and want your work to matter to real families across Southeast Asia, this role is for you.

ABOUT TEENCARE AI

At TeenCare AI, we work at the intersection of education, family, and technology. Each week, teens work with a dedicated 1:1 mentor. These sessions generate rich behavioral signals that our AI translates into structured intelligence for parents — helping families make better decisions at the moments that actually matter.

We combine human mentorship with AI analysis. Our platform is already used by families across Southeast Asia, and we are building toward a system that can scale globally.

WHAT YOU’LL DO

Build the AI layer of the product

• Design and maintain RAG pipelines that ground behavioral insights in each family‘s history
• Integrate and orchestrate LLMs (OpenAI, Google, and others) for analysis, summarization, and structured output
• Build and iterate on prompt systems — structured prompting, few-shot examples, and evaluation frameworks
• Develop AI-powered features in the Parent Copilot and mentor tooling

Own the AI infrastructure

• Design and maintain data pipelines that feed behavioral models
• Build evaluation pipelines to measure model accuracy and output quality over time
• Manage vector databases and embedding workflows for retrieval systems
• Monitor AI system performance and reduce latency and cost

Collaborate closely with the full-stack engineer

• Integrate AI features into the product end-to-end — from model to UI
• Work together on streaming AI responses (SSE) and real-time AI interactions via WebSockets
• Contribute to backend services where AI intersects with product logic

Experiment and improve

• Run structured experiments on model performance, prompt variants, and retrieval strategies
• Explore fine-tuning opportunities as our dataset of teen-parent interactions grows
• Stay sharp on what is happening in applied AI and bring relevant advances back to the team

AI STACK & ENGINEERING EXPECTATIONS

LLM Platforms & Orchestration

• OpenAI API
• Anthropic API
• LangChain / LlamaIndex
• Prompt Engineering

RAG & Vector Systems

• Pinecone / Weaviate
• pgvector
• Embeddings
• Chunking Strategies

ML & Fine-Tuning

• Python
• HuggingFace
• Fine-tuning (LoRA)
• RLHF Basics

AI Infrastructure & Pipelines

• Data Pipelines
• Eval Frameworks
• SSE / Streaming
• MLflow / W&B

Collaborative Stack (shared with fullstack engineer)

• Node.js
• PostgreSQL
• MongoDB
• Redis

YOUR SKILLS & EXPERIENCE

Technical

• 1–2 years of hands-on experience building with LLMs in a product or research context
• Solid Python skills — you are comfortable building pipelines, not just notebooks
• Working knowledge of RAG architectures and vector retrieval
• Experience integrating LLM APIs and managing prompt complexity at scale
• Familiarity with evaluation methods for generative AI outputs
• Basic awareness of model fine-tuning workflows (even if you have not done it end-to-end yet)

Mindset

• Genuinely curious about applied AI — you experiment, you read, you iterate
• Comfortable with ambiguity — you can make reasonable decisions without a full spec
• Strong ownership — you follow through on your work from experiment to production
• Collaborative by default — you work well paired with a fullstack engineer and share context freely
• Care about the outcome — you want your AI work to actually help families, not just pass benchmarks

WHAT SUCCESS LOOKS LIKE

Within the first 3 months, you will:

• Ship production AI features used by real families every week
• Build and maintain a reliable RAG pipeline powering the Parent Copilot
• Own a prompt evaluation framework that catches regressions before they reach users
• Reduce AI output errors measurably with structured testing and iteration
• Contribute meaningfully to TeenCare’s behavioral intelligence architecture

WHY YOU’LL LOVE WORKING HERE

• Work directly with the Founders, CEO, and Leadership Team
• Competitive salary, negotiable based on ability, with a performance bonus
• Transparent salary and level increase roadmap — Performance Review every 3 months
• Full social insurance, health insurance, and Company benefits in accordance with the Labor Law: 12 days of paid leave/year, work from home flexibility, 90 minutes of menstrual leave/month
• Strong learning culture: Onboarding, Monthly Learning & Sharing sessions, Builder Forum, and a personal Learning Hub
• Internal bonding: weekly Happy Hour, quarterly Team Building, Year End Party, and a rotating Pantry menu
• A young, ambitious, and deeply human team united by the mission to create real value for children and families

WORKING HOURS

Monday – Friday + Saturday morning

8:30 AM – 12:00 PM

1:30 PM – 6:00 PM

TeenCare AI is building the future of family intelligence in Southeast Asia.

We’re a team that ships, learns, and cares deeply about the families we serve.