Senior Machine Learning Engineer

About the role

A senior role steering machine learning strategy and building production LLM systems for social connection products. The work covers model training, agent tooling and evaluation that shape user experience and revenue.

What you will do

• Set the machine learning roadmap and connect each strategic bet to concrete business goals
• Own LLM post-training from dataset building through fine-tuning and preference optimisation
• Create agent harnesses and systems where LLMs call tools and use structured outputs
• Design evaluation frameworks and automated judge systems that predict real-world quality
• Improve conversation quality in memory, emotional intelligence, persona stability, safety and loop avoidance
• Deliver production systems that influence retention, monetisation and the user experience

What they look for

• Expert Python skills covering clean code, typing, async, multiprocessing, profiling and debugging
• Hands-on transformer work, LLM training and fine-tuning with SFT, LoRA, QLoRA and PEFT
• Preference optimisation methods such as DPO, ORPO, SimPO, CPO or GRPO, or readiness to learn them fast
• Dataset construction for training: cleaning, filtering, deduplication, preference pairs, synthetic data and hard negatives
• LLM evaluation experience including offline and pairwise evals, judges and human annotation
• Agent harness design with tool calling, routing, retries, state, memory, guardrails and observability

Nice to have

• Experience with computer vision, image generation or multimodal work
• Long-running conversations and character consistency
• Serving stacks such as vLLM, TGI or SGLang
• DeepSpeed, FSDP or Accelerate, quantization and safety classifiers

Benefits

• Full-time remote work
• 28 calendar days of vacation each year
• 7 wellness days annually for personal matters
• Referral bonuses of up to $5000
• Half of professional training, conference and meeting costs covered