Quantitative Researcher

Quantitative Researcher | World Models & Quantitative Perception

Astera Holdings
Remote

About Astera

Astera is building decision intelligence for events across markets. Our systems transform noisy real-world events into structured, actionable intelligence across sports, prediction markets, macro, crypto, and equities.

We are building toward generalized world models capable of understanding dynamic environments, extracting latent structure from partially observed systems, and improving decision quality under uncertainty.

We are pursuing problems at the intersection of:

• world models,
• multimodal reasoning,
• computer vision,
• quantitative inference,
• agentic systems,
• and event-driven intelligence architectures.

The Role

We are hiring a Research Engineer focused on world models, quantitative perception systems, and latent-state reasoning.

You will work on systems that:

• model dynamic environments in latent space,
• extract actionable signal from noisy multimodal data,
• infer hidden state transitions,
• and quantify qualitative phenomena into structured representations usable by downstream agents and decision systems.

This role sits between:

• applied research,
• quantitative modeling,
• computer vision,
• reinforcement learning,
• and systems engineering.

You should be comfortable operating in ambiguous, frontier-style research environments with a high degree of autonomy.

Responsibilities

• Develop latent-space and world-model architectures for dynamic real-world systems
• Build models that infer hidden state from noisy or partially observed environments
• Design quantitative frameworks for extracting signal from high-dimensional data
• Research and implement multimodal reasoning systems across vision, temporal, and structured data
• Build spatiotemporal perception and forecasting pipelines
• Develop representation-learning systems for event understanding and state estimation
• Design agent memory and long-horizon reasoning mechanisms
• Build research-grade experimentation, evaluation, and simulation frameworks
• Collaborate with infrastructure, AI, and product teams to productionize research systems
• Translate qualitative real-world phenomena into measurable quantitative abstractions

Qualifications

Required

• Strong background in machine learning, applied mathematics, computer science, physics, quantitative research, or a related technical field
• Experience building ML systems in Python using PyTorch, JAX, or TensorFlow
• Experience modeling complex systems under uncertainty
• Strong understanding of probabilistic reasoning and statistical inference
• Experience working with noisy, high-dimensional, or partially observed datasets
• Ability to independently design and run research experiments
• Strong systems-thinking and problem-solving ability

Preferred Experience

World Models & Latent-State Modeling

Experience with:

• latent-space modeling,
• predictive world models,
• sequence modeling,
• state-space models,
• memory architectures,
• reinforcement learning,
• trajectory modeling,
• or agent-based systems.

Computer Vision & Perception

Experience with:

• video understanding,
• object tracking,
• spatiotemporal forecasting,
• multimodal perception,
• vision transformers,
• segmentation/detection systems,
• sports tracking,
• robotics perception,
• or sensor fusion systems.

Quantitative Signal Extraction

Experience:

• extracting signal from noisy environments,
• identifying weak predictive structure,
• designing inference systems under uncertainty,
• or converting qualitative observations into quantitative representations.

Physics-Based & Causal Modeling

Experience with:

• dynamical systems,
• simulation environments,
• physics-informed ML,
• causal inference,
• or state transition modeling.

Technical Stack

We value strong engineering fundamentals more than specific tools, but experience with the following is highly relevant:

• Python
• PyTorch
• JAX
• TensorFlow
• C++
• CUDA
• OpenCV
• RL frameworks
• Distributed training systems
• Scientific computing libraries
• Time-series and probabilistic modeling frameworks

What We Look For

• High intellectual rigor
• Strong research intuition
• Systems-level thinking
• Comfort operating under ambiguity
• Fast iteration velocity
• Curiosity across domains
• Strong quantitative reasoning
• Bias toward truth-seeking over consensus
• Ability to extract structure from disorder

We are specifically interested in people capable of quantifying the qualitative.

Nice-to-Have Backgrounds

• Autonomous systems
• Robotics
• Quantitative trading
• Scientific computing
• Aerospace / space systems
• Sports analytics
• Knowledge graphs
• Agentic systems
• Real-time inference systems
• High-performance ML infrastructure

Compensation

• Competitive salary
• Meaningful equity participation
• Opportunity to work on frontier-scale problems with a highly technical team
   

NYC. 

• Resume / LinkedIn
• GitHub or research portfolio 
• Representative projects, papers, or systems you’ve built
  info@astera.holdings
  
  Ad Astra