AI Software Engineer (AI-Native Development)

3. WHAT YOU WILL OWN

A. Build AI-Powered Products

• Build applications where AI is a core capability.
• Integrate AI models and APIs into production software.
• Develop AI-assisted user experiences.
• Turn prototypes into maintainable systems.
• Evaluate whether AI is actually the right solution before implementing it.

B. Build LLM & Multimodal Features

Work with capabilities such as:

• Large Language Models
• Multimodal AI
• Structured Outputs
• Tool / Function Calling
• Embeddings
• Semantic Search
• Retrieval-Augmented Generation (RAG)
• AI Agents
• Image, Video, Audio or Text AI services where required

Choose solutions based on quality, reliability, latency, cost and business value.

C. AI Agents & Automation

• Identify repeatable processes that can be automated.
• Build agentic and workflow-based systems.
• Connect AI with APIs, databases and business tools.
• Implement human review where full automation is inappropriate.
• Measure whether automation improves speed, quality, productivity or cost.

D. Build Complete Software Around AI

AI alone is not a complete product.

You should be capable of building or contributing to:

Frontend • Backend • APIs • Databases • Authentication • Integrations • Cloud Deployment • Monitoring

Build complete systems where AI works as part of a reliable software architecture.

E. Reliability, Security & Guardrails

Design for risks including:

• Hallucination
• Incorrect Outputs
• Prompt Injection
• Data Privacy
• Security
• Model Failure
• Latency
• API Dependency
• Cost
• Inconsistent Behaviour

Implement validation, guardrails, fallbacks, monitoring and human review where required.

F. AI Evaluation

• Define success criteria before deployment.
• Build appropriate test and evaluation processes.
• Compare prompts, models and architectures.
• Measure quality, accuracy, latency and cost.
• Evaluate failures instead of only successful examples.
• Use evidence when deciding whether an AI feature is ready to scale.

G. AI-Native Development Workflow

Use AI effectively across:

Coding • Debugging • Testing • Refactoring • Documentation • Research • Prototyping • Code Review

AI should make you faster, but not careless.

You remain responsible for understanding and validating what is shipped.

H. Explore New AI Opportunities

• Stay current with rapidly changing AI technology.
• Prototype promising capabilities quickly.
• Separate useful technologies from hype.
• Recommend what ToYm should adopt, test, delay or avoid.
• Convert successful experiments into production systems.

4. END-TO-END OWNERSHIP

You may work with:

Product • Technology • UI/UX • Growth • Content • Operations • Management • Other Developers

You should be able to move through:

Understand Problem → Determine Whether AI Helps → Design → Prototype → Evaluate → Build → Deploy → Monitor → Improve

Your responsibility is not:

“The model returned an answer.”

Your responsibility is:

“The complete system produces a useful and reliable outcome.”

5. QKSC — WHO WE ARE LOOKING FOR

We assess candidates through:

Q — Qualification
Do you have evidence of building real AI-powered software?

K — Knowledge
Do you understand how modern AI systems actually work?

S — Skill
Can you build and deploy them?

C — Competency
Can you repeatedly turn uncertain technology into useful results?

Our hiring approach follows:

Ask → Evidence → Scenario → Practical Test → Reference Check

6. Q — QUALIFICATION

There is no strict degree requirement and no fixed minimum number of years.

You must, however, demonstrate real evidence of AI development.

You should have personally built or significantly contributed to at least one:

• AI-Powered Application
• LLM Feature
• AI Agent
• RAG System
• AI Automation
• AI-Integrated Software Product
• Other Meaningful AI System

Where available, provide GitHub, live systems, demonstrations, architecture examples, portfolio work or other verifiable evidence.

We want to understand:

What did YOU personally build?

A working system with clear technical thinking is more valuable than a long list of AI tools on a CV.

7. K — KNOWLEDGE

You should demonstrate practical understanding of areas such as:

LLMs • Prompt & Context Design • Structured Outputs • Tool Calling • Agents • RAG • Embeddings • Vector Search • Model Selection • AI APIs • Evaluation • Hallucination Management • Guardrails • Security • Prompt Injection • Token / API Cost • Latency • Context Windows • Data Privacy • Fallback Strategies • Monitoring

You should also understand software engineering fundamentals involving:

Backend Development • APIs • Databases • Authentication • Git • Testing • Deployment • Cloud Infrastructure • System Integration

You do not need to know every model or framework.

AI technology changes quickly.

We care more about whether you understand the fundamentals and can adapt.

8. S — SKILL

You should be able to:

• Build an AI feature from idea to production.
• Integrate AI models through APIs.
• Build RAG or retrieval workflows where appropriate.
• Implement structured outputs and tool calling.
• Build agentic or automated workflows.
• Connect AI systems to databases and external services.
• Evaluate model output systematically.
• Control reliability, latency and cost.
• Build appropriate fallbacks and guardrails.
• Debug AI and conventional software problems.
• Deploy and monitor AI-powered applications.
• Use AI to improve your own development productivity.
• Explain why you selected a particular model or architecture.

9. C — COMPETENCY

We particularly value:

Ownership — You take responsibility for the complete outcome.

Problem Solving — You start with the problem, not with an AI tool.

Experimentation — You test ideas quickly and learn from evidence.

Effective Execution — You convert prototypes into usable software.

Product Thinking — You understand whether AI creates genuine user or business value.

Critical Judgement — You recognise when AI should not be used.

Learning Agility — You can continuously adapt as models and tools change.

Responsibility — You consider privacy, security, reliability and cost.

Communication — You can explain AI concepts clearly to non-AI specialists.

Working Under Ambiguity — You can explore uncertain problems and progressively determine the right solution.

10. WHAT MATTERS MORE THAN YEARS OR CERTIFICATES

We want to understand:

• What AI systems have you built?
• What problem did they solve?
• What did you personally develop?
• Which model or architecture did you choose and why?
• How did you evaluate quality?
• What happened when the AI produced incorrect results?
• How did you manage cost and latency?
• Was it deployed to real users?
• What measurable result did it create?
• Can you move beyond a prototype into production?
• What would you change if usage increased significantly?

11. BITESHORTS & TOYM CONTEXT

You may work on AI opportunities across BiteShorts and other ToYm products.

Potential areas may include:

Content Discovery • Search • Personalisation • Recommendation • Content Operations • Creator Tools • Merchant Tools • Advertising • Customer Support • Analytics • Workflow Automation • Internal Productivity • New AI-First Products

We do not want to add AI simply because AI is popular.

Every implementation should answer:

What problem are we solving?
Why is AI appropriate?
How do we measure whether it works?

12. SELECTION PROCESS

Shortlisted candidates may be asked to:

• Demonstrate an AI system they personally built.
• Explain its architecture and model selection.
• Explain how they evaluated output quality.
• Diagnose a scenario involving hallucination, latency, cost or reliability.
• Design an AI solution for a real ToYm or BiteShorts use case.
• Complete a practical assessment involving AI integration, RAG, agents, APIs, debugging or system design.

We may also ask:

If the prototype works for 100 users, what changes when it must support 100,000?

We want to understand both your AI capability and engineering judgement.

13. OUR VALUES

CUSTOMER-CENTRIC
PURPOSE-DRIVEN & RESULTS-ORIENTED
TEAMWORK
RESPONSIBILITY
EFFECTIVE EXECUTION
HUMBLE & LEARNING
PURSUIT OF EXCELLENCE
CONTINUOUS INNOVATION
CONTRIBUTE TO SOCIETY

14. THIS ROLE IS FOR YOU IF…

You already use AI to change how software is built.

But you understand that being AI-native means more than prompting a model.

You can combine AI capability + software engineering + product thinking to build systems that actually work.

You experiment quickly, evaluate objectively and strengthen what succeeds before putting it into production.

And you want to help build AI capabilities across BiteShorts and ToYm’s growing digital products.

To Apply: Submit your resume and, where available, GitHub, technical portfolio or examples of AI applications, agents, automations or systems you personally built. Be prepared to explain the problem, architecture, model choice, personal contribution, evaluation method and result.