← Back to All Openings

Lead Systems Engineer – Dynamic Belief‑Graph Regularization

corpora-jobs-1778796293285-db9d41c6 - Frontier Development
Systems EngineerLead1 position
⚡

Why This Role is Different

Frontier Development Role

Own the end‑to‑end design of a graph‑based belief regularizer that keeps multi‑agent reasoning robust to deceptive inputs while staying computationally efficient enough for real‑time deployment.

The Frontier Element

You will pioneer a dynamic, non‑monotonic belief graph that simultaneously tracks credibility, confidence, and structural support, and enforce it through a lightweight regularizer integrated into a generalized multi‑relational GCN.

🔬

Project Context

Research Area

Dynamic Belief‑Graph Regularization (DBGR)

From: Theory of Mind Defenses Against Communication Sabotage

Why This Role is Critical

Implement the graph‑based soft constraint that limits the influence of malicious messages on belief updates, ensuring local coherence and preventing catastrophic drift.

What You Will Build

A production‑grade belief‑graph inference engine built on GEM‑GCN, runtime regularization module, and a benchmarking suite for belief consistency and latency.

🛠

Key Responsibilities

  • Design and implement the GEM‑GCN‑based belief‑graph inference engine with dynamic edge‑type handling.
  • Integrate the DBGR regularizer into the agent’s belief update loop, ensuring sub‑10 ms inference latency.
  • Develop a suite of synthetic and real‑world benchmarks to evaluate belief drift, consistency, and robustness.
  • Collaborate with the AC‑ToM team to align graph semantics with adversarial scenarios.
  • Document and open‑source the core library for internal use and community adoption.
🎯

Required Skills & Experience

Technical Must-Haves

Graph Neural Networks (GNNs)

Expert
Implementing GEM‑GCN and dynamic message passing.

Causal inference & non‑monotonic reasoning

Advanced
Designing credibility‑confidence semantics.

Distributed systems & low‑latency inference

Advanced
Ensuring real‑time performance across agents.

C++/CUDA programming

Expert
Optimizing graph operations for sub‑millisecond latency.

Python, PyTorch

Proficient
Rapid prototyping and integration.

Experience Requirements

  • 6+ years in graph ML or related systems engineering.
  • 3+ years in production‑grade, low‑latency inference.
  • Published work on graph regularization or belief‑graph models.

Education

PhD in Computer Science, AI, or Electrical Engineering.

⭐

Preferred Skills

  • Experience with meta‑learning for graph models.
  • Hardware acceleration (FPGA/ASIC) for GNN inference.
  • Knowledge of TSLink or similar low‑latency DSP architectures.
🤝

You Will Thrive Here If...

  • Owns end‑to‑end system design and delivery.
  • Comfortable with ambiguity and rapid prototyping.
  • Values measurable impact and continuous improvement.
📈

Impact & Growth

12-Month Impact

Achieve a 40% reduction in belief drift under malicious messages, maintain sub‑10 ms inference latency, and publish a case study demonstrating DBGR’s effectiveness in a real‑time IoT setting.

Growth Opportunity

Architect the multi‑agent belief system across the company’s product portfolio, mentor junior engineers, and lead cross‑functional integration with the TTVL team.

Ready to Push the Boundaries?

If this sounds like the challenge you have been looking for, we want to hear from you. We value what you can build over where you have been.