[全遠端日本新創 ] Founding AI Engineer ( Potential Lead)

A pioneering technology startup seeks a Founding Senior AI Engineer to build and own the AI layer of an industrial enterprise platform using a hybrid self-hosted stack with GPU access and APAC remote flexibility.

Key responsibilities:

As a Founding Senior AI Engineer, you will build and deploy self-hosted LLM and vision systems for an industrial AI platform, managing GPU infrastructure and collaborating across Tokyo and Taipei teams.

Model Development & Fine-Tuning

Deploy self-hosted LLMs (Llama 3, Mistral, Code Llama) on GCP GPU infrastructure
Fine-tune vision models (e.g., LLaVA, Qwen-VL, Florence-2) for P&ID and engineering drawings
Build and maintain data pipelines for industrial documents (P&ID, PLC), including annotation and versioning
Implement 3-layer P&ID extraction (vector text, tile-based vision, ISA 5.1 validation)
Apply LoRA/QLoRA/PEFT to improve document understanding accuracy

AI Infrastructure & MLOps

Operate production inference stacks (vLLM, TGI, Ollama) on GCP
Manage H100/A100 GPU workloads on Vertex AI for scalable serving
Benchmark models vs frontier APIs (accuracy, latency, cost) with evaluation pipelines
Build domain RAG systems using ISA 5.1, IEC 62443, IEC 61131-3 standards
Own evaluation, regression testing, and continuous monitoring frameworks

Platform & Product Integration

Integrate inference services with Tokyo backend (AIDC APIs)
Support Taipei solutions team with customer demos and performance benchmarks
Translate fine-tuning insights into product roadmap priorities
Maintain documentation on model capabilities, limits, and failure modes

Candidate profile:

To excel in this role, you bring experience deploying ML models, fine-tuning LLMs/vision systems, building inference infrastructure, strong Python skills, and English/Mandarin collaboration ability.

4+ years’ ML engineering with production model deployment experience (not research-only or pure data science)
Hands-on LLM fine-tuning (LoRA, QLoRA, PEFT) on real-world datasets with measurable outputs
Experience with vision/document models (e.g., LLaVA, Qwen-VL, Florence-2 or equivalent)
Production inference systems (vLLM, TGI, Ollama) with ability to meet latency/SLA requirements
GCP Vertex AI or equivalent experience managing GPU clusters (H100/A100 orchestration)
Strong Python skills (PyTorch, Transformers, HuggingFace) with ability to implement ML papers
Strong English communication; Mandarin Chinese highly preferred

Nice to have

Domain fine-tuning on industrial/engineering data (P&ID drawings, SCADA/PLC code, instrumentation documents)
Experience with RLHF or DPO to improve model quality using structured human feedback
Computer vision for document understanding (layout analysis, OCR, table and symbol extraction)
Familiarity with industrial standards such as IEC 62443 (OT security), ISA 5.1 (P&ID symbology), or IEC 61131-3 (PLC programming)
Prior experience in an AI startup or embedded AI research team within an enterprise environment

About the company:

A diversified global organisation operates across data-driven technology and industrial solutions, delivering scalable platforms and measurable impact in complex enterprise environments. It fosters a collaborative, engineering-led culture focused on innovation, long-term growth, and enabling teams to solve high-value problems with autonomy and modern tools.

Keywords: AI solutions, Smart Space, SaaS, enterprise automation, self-hosted model stack, fine-tuning pipelines, inference infrastructure, start-up tech firm

What's next:

Build what matters—apply now and shape the future of intelligent systems.

相似職缺

瀏覽更多職缺

[全遠端日本新創 ] Founding AI Engineer ( Potential Lead)

分享

相似職缺