AI/ML Engineer

Role Overview

We are seeking an AI/ML Engineer to design, implement, and optimize the machine learning systems that power HoloMe's holographic avatars. This role focuses on natural language understanding, speech-to-text (STT), text-to-speech (TTS), knowledge base integration, and real-time inference, ensuring our avatars deliver natural, responsive, and localized interactions across retail, transit, and event deployments.

______________

Key Responsibilities

Core AI & NLP Development

Integrate and fine-tune LLMs (OpenAI, Hugging Face, or custom models) for contextual, domain-specific responses.

Implement and optimize Retrieval-Augmented Generation (RAG) pipelines using PostgreSQL + pgvector for semantic search

Byondasia-Technical Infrastruct

Develop and maintain conversational flows that align with product and compliance requirements.

Speech & Voice Systems

Implement speech-to-text pipelines (Whisper, Azure STT, Google Speech API) for real-time transcription.

Integrate and optimize TTS pipelines (Klleon, ElevenLabs, Azure Cognitive Services) for natural, localized voice synthesis.

Support voice cloning and persona-based vocal delivery for branded avatars.

Knowledge & Context Management

Structure knowledge bases (JSON/CSV/DB) for domain-specific queries (e.g., metro stations, product SKUs, event FAQs).

Implement multi-language support (Malay, English, Arabic, Chinese, Japanese).

Ensure semantic accuracy by testing embeddings, tagging, and metadata management.

Optimization & Deployment

Optimize inference pipelines for low-latency performance on cloud GPU clusters and edge devices (Holobox PCs).

Build fallback logic for offline/local responses when cloud connectivity is disrupted.

Work with Cloud & DevOps engineers to scale GPU workloads in Kubernetes clusters.

Testing & Monitoring

Develop test harnesses for AI interactions (mock queries, edge case handling, stress tests).

Collaborate with QA to automate conversation validation and accuracy benchmarking.

Monitor model drift, bias, and performance degradation across deployments.

Compliance & Data Protection

Ensure no intentional collection of personal data (names, IDs, ages, etc.).

Apply anonymization, logging rules, and retention policies aligned with PDPL/GDPR/CCPA.

Work with compliance and DevOps to audit AI model accuracy, security, and localization

______________

Qualifications

Must-Have Skills

Strong experience in NLP, LLM fine-tuning, and RAG pipelines.

Proficiency with Python (PyTorch, TensorFlow, Hugging Face Transformers).

Knowledge of embeddings, semantic search (pgvector, FAISS, or Pinecone).

Hands-on experience with STT and TTS systems.

Familiarity with cloud-based ML deployment (AWS, Azure, or Kubernetes).

Nice-to-Have Skills

Experience with multi-language NLP (especially Malay, Arabic, Chinese, Japanese).

Real-time/streaming ML experience (ASGI, WebSockets, gRPC).

Familiarity with Redis + Celery for async workflows.

Experience building chatbot or voice assistant pipelines.

Knowledge of security practices in AI (prompt injection defense, logging sanitization).