Senior AI Systems & Time-Series Engineer (Global - Remote)

Job Description

Contract-to-hire
Partial timezone overlap: PST 6am-12pm (flexible)

We are seeking a Senior AI Systems & Time-Series Engineer with deep expertise in probabilistic modeling, time-series forecasting, and modern AI architectures such as JEPA and reinforcement learning. This role is ideal for someone who thrives in noisy, high-uncertainty domains such as finance or space operations and can drive production-grade deployment of AI systems. You will work on a small, high-ownership team focused on building real, end-to-end systems — not just experiments. Remote-friendly, with significant overlap in US Eastern time preferred.

Core Mathematical & Modeling Skills

• Deep probability theory and statistics
• Strong time-series experience in noisy, high-uncertainty domains (finance/crypto strongly preferred; space or industrial operations also valued)
• Comfortable with uncertainty quantification, false positives, noisy signals, proper evaluation metrics

Methodologies & Architectures

Hands-on building and training of:

• Joint Embedding Predictive Architectures (JEPA-style)
• Reinforcement Learning (offline RL, Decision Transformers, etc.)
• Modern transformers for sequence/decision modeling
• Bayesian/probabilistic methods and uncertainty estimation experience

Explicitly looking for real modeling experience, not just “calling someone else’s LLM API”

Systems, Deployment & Data Pipelines

• Containerized model deployment and orchestration (Singularity, Apptainer, multi-node)
• Building and maintaining large-scale training/inference pipelines
• Experience with HPC/supercomputing environments (job submission, large datasets, node-hour scaling) is a strong plus

Application Domains

• Trading / financial time-series
• Space operations and industrial facility monitoring
• Hierarchical agentic systems and workflow automation

General Profile

• Ships production-grade systems, not just research notebooks
• Can produce and present tangible evidence of progress (plots, graphs, statistics, MLflow dashboards, etc.) to highly technical stakeholders
• Small team, high ownership, remote-friendly (significant US Eastern time overlap)