Qube Media Hiring! Full Time AI Agent - Prompt Engineer in Federal Territory - Ricebowl

The Al Agent & Prompt Engineer will design, build, evaluate, deploy, and improve LLM-powered agents used in customer-service and operational workflows.

The role combines prompt and conversation design with software integration, testing, analytics, and production ownership.

You will work closely with Qube Media's CTO, Product/Project Manager, developers, and client-facing teams to turn real service problems into dependable Al workflows. The priority is not experimentation alone: successful prototypes must be measurable, supportable, secure, and capable of moving into production.

Key Responsibilities

• Build, maintain, and improve production conversational Al agents across Quins Al Voice and Quins Al Chat.

• Own system prompts, prompt templates, tool and function-calling logic, conversation flows, fallbacks, escalation paths, and response policies.

• Translate customer-service and operational problems into agentic workflows by decomposing tasks, defining decision logic, and connecting the required systems.

• Develop retrieval-augmented generation (RAG) pipelines using approved knowledge bases, document stores, databases, and vector-search components.

• Integrate agents with REST APIs, CRM systems, telephony services, WhatsApp, web channels, spreadsheets, databases, analytics tools, and other approved platforms.

• Design multilingual evaluation datasets and test cases for English, Bahasa Malaysia, and Malaysian conversational contexts, including code-switching where relevant.

• Measure and improve answer quality, task completion, containment, escalation accuracy, hallucination rate, latency, reliability, and operating cost.

•Prototype proofs of concept with Product, Operations, and client stakeholders, then convert validated use cases into maintainable production components.

• Monitor live agents, investigate failures, analyse logs, reproduce issues, and implement reliability, safety, privacy, and guardrail improvements.

• Support the integration of MYEMOSI capabilities where emotional or behavioural signals are relevant to routing, prioritisation, analytics, or response adaptation.

• Maintain technical documentation, prompt and configuration versioning, evaluation records, release notes, decision logs, and handover materials.

• Keep source code, prompts, datasets, documentation, and product components within Qube Media-approved repositories and governance processes to protect product ownership and intellectual property.

Required Qualifications

• Bachelor's degree in Computer Science, Artificial Intelligence, Software Engineering, Data Science, or a related discipline, or equivalent demonstrated capability.

• Strong Python skills and practical experience building API-based integrations or backend services.

• Hands-on experience with LLM APls and at least two of the following: prompt engineering, RAG, tool/function-calling, structured outputs, agent frameworks, or conversational workflow orchestration.

• Ability to build and test an end-to-end working prototype, not only notebooks or isolated model experiments.

• Working knowledge of Git, SQL, API authentication, error handling, testing, and software-development fundamentals.

• Evidence of delivery through production work, internships, hackathons, university projects, GitHub repositories, demonstrations, or technical write-ups.

• Analytical approach to experimentation, evaluation, root-cause analysis, and iterative improvement.

• Clear written and verbal communication, with the ability to explain technical decisions to non-technical stakeholders.

• Comfort working in a fast-moving environment where requirements may evolve as customer and product knowledge improves.

• Professional proficiency in English and Bahasa Malaysia.

Experience Level

Applicants with approximately one to three years of relevant experience are preferred. Strong fresh graduates may also be considered when they can demonstrate a credible LLM-agent portfolio, solid software fundamentals, and the ability to learn and ship quickly.

Preferred / Nice-to-Have Experience

• Customer-service chatbots, voicebots, contact-centre automation, CRM workflows, or case-management systems.

• Speech-to-text, text-to-speech, telephony, real-time audio, speech analytics, or speech-emotion recognition.

• WhatsApp Business Platform, web chat, messaging integrations, or communications platform-as-a-service tools.

• Malaysian multilingual data, local accents, Bahasa Malaysia-English code-switching, Manglish, or other

Malaysian languages.

• LLM evaluation frameworks, prompt testing, tracing, observability, red-teaming, or guardrail tooling.

• FastAPI, PostgreSQL, Docker, vector databases, cloud deployment, CI/CD, or production monitoring.

• Data visualisation, KPI dashboards, customer-behaviour analytics, or operational reporting.

• Privacy, security, responsible-Al controls, accessibility, or Malaysian PDPA considerations.

• Current resume or curriculum vitae.

• GitHub profile, portfolio, demonstration links, or technical write-ups showing relevant work.

• A short explanation of one Al agent, automation, or software project you built, including your specific contribution, the architecture, how you evaluated it, and what you would improve next.