.png){kind=logo}
Siliang Zhang
Intern Machine Learning Engineer specializing in LLMs, RAG, and vision-language systems
Connect with Siliang
Siliang already has a relationship with Reval, so a warm intro from us gets a much better response than cold outreach.
Recommended
Already have an account?
About
Robotics ML/software engineer focused on Vision-Language-Action control for 7-DoF robots, replacing tokenized action decoding with continuous regression heads (including a logit-weighted expectation approach) to improve stability and real-time behavior. Strong in ROS1/ROS2 systems integration and debugging closed-loop manipulation issues via latency instrumentation, QoS-aware distributed messaging, and sim-to-real validation using Gazebo/Unity, Docker, and CI pipelines.
Experience
Education
Key Strengths
- Designed and implemented continuous regression-based action decoding for VLA robot control (replacing token-based decoding)
- Improved training stability for continuous actions via Huber loss tuning, gradient clipping, and entropy regularization
- Diagnosed real-time oscillation/jitter by instrumenting end-to-end latency and control loop variance with timestamped logs
- Reduced inference latency and jitter by eliminating autoregressive token loops and optimizing ROS execution architecture
- Integrated scale-aware numeric embeddings (xVal, FoNE) and validated impact with ablations for better OOD robustness
- Built ROS1/ROS2 inference and logging nodes with synchronization, preprocessing, and safety constraints
- Designed resilient distributed-robot communication using ROS2 DDS QoS, asynchronous coordination, and heartbeat/health checks
- Established reproducible sim-to-real workflows using Gazebo/Unity simulation plus Docker packaging and GitHub CI pipelines
Discover more candidates like Siliang
Search across thousands of pre-screened, high-quality, high-intent candidates on Reval.
Search Talent