An early-stage AI company building a next-generation personal assistant designed to handle everyday tasks — communication, scheduling, organizing, follow-ups — with little to no user input required. The core engineering challenge is reliability: getting AI systems to execute multi-step, long-running workflows consistently, even when the underlying models behave unpredictably. The product aims to meaningfully cut down the time people spend on daily admin and coordination.
The Role
A senior, individual-contributor ML engineering position with full ownership of key production ML systems. This person will take vague, open-ended problems and turn them into working, scalable solutions — not a research-only role, but one grounded in shipping and maintaining live systems.
What You'll Do
- Design and build the ML infrastructure behind a long-running, proactive AI product
- Own the full lifecycle — data, training, evaluation, inference, deployment, and ongoing tuning
- Convert experimental/research concepts into dependable production systems
- Diagnose and fix model and pipeline issues using live production data
- Work in fast iteration cycles — release, measure, adjust, repeat
- Partner closely with research, product, and engineering counterparts
- Provide technical mentorship and code/design review to other ML engineers
- Balance competing constraints: latency, infrastructure cost, reliability, and safety
Stack
Python, PyTorch/JAX, GPU-based training and inference infrastructure
What We're Looking For
- Track record of shipping ML systems that real users depend on
- Strong intuition for how ML models fail in the real world, not just in theory
- Systems-level thinking, not just scripting — clean, production-grade code
- High autonomy — comfortable owning problems without close direction
- Fast learner, clear communicator, iterates well on feedback
Success Looks Like
- Production ML systems hitting targets for accuracy, latency, cost, and reliability
- Fast diagnosis and resolution of production issues, minimal user-facing disruption
- Pipelines (training/inference/data) that scale and hold up over time
- Visible, measurable improvements driven by real usage data
- Peers leveling up through your review and mentorship
- ML work integrating smoothly into the broader product
Team Culture
Small, high-caliber team, flat decision-making, fast pace. Expect autonomy and structure to coexist — you're trusted to self-direct, but expected to bring rigor.
ST
Reg No. R1768414
BeathChapman Pte Ltd
Licence no. 16S8112
