Forward Deployed Engineer – Inference, Post-Training, Mandarin Speaking

This is a fully remote position, open to applicants in Singapore.

📋 Description

• Select, configure, and enhance inference engines based on hardware specifications, model architecture, and workload profiles.

• Develop configuration modifications to secure critical proofs of concept (POCs) and benchmarks while optimizing customer deployments.

• Fine-tune KV cache, implement speculative decoding, assess optimal tensor and pipeline parallelism, and devise quantization strategies to achieve throughput and latency objectives.

• Lead hands-on reinforcement learning (RL) training runs and refine system design.

• Assist customers through LoRA, SFT, DPO, RLHF, and GRPO pipelines from experimentation to production.

• Serve as the primary technical contact for strategic accounts.

• Monitor and refine endpoint configurations to help customers maximize platform value.

• Collaborate with customers to reach critical milestones.

• Establish direct alignment with strategic customers during the onboarding process.

• Ensure that proper inference and post-training configurations are implemented from day one.

• Provide field insights to influence the software and model roadmap.

• Contribute to product enhancements to meet customer needs and improve user experience.

• Promote early adoption of features and research initiatives with strategic customers.

⛳️ Requirements

• Must be a permanent resident or citizen of Singapore.

• Over 5 years of experience in a technical role, focusing on inference systems, open-source LLM deployment, or post-training workflows.

• Expert-level, hands-on experience with inference engines such as vLLM, TensorRT-LLM, or SGLang.

• Ability to identify and resolve performance issues at the engine level.

• Extensive knowledge of KV cache tuning, speculative decoding, tensor parallelism, pipeline parallelism, and quantization techniques.

• Practical experience with fine-tuning and post-training pipelines, including LoRA, SFT, DPO, RLHF, and GRPO.

• Capability to advise on system design.

• Comprehensive knowledge of leading open-source models.

• Strong judgment regarding model selection for specific customer use cases, hardware profiles, and performance goals.

• Proficient in Python programming.

• Comfortable operating in production environments.

🏝️ Benefits

• Competitive salary.

• Equity in a startup.

• Health insurance coverage.

• Additional benefits.

• Flexibility regarding remote work options.