Scientific Computing Associate - VR Technologies for Social Neuroscience

Howard Hughes Medical Institute•Ashburn, VA

1d•Onsite

About The Position

The Scientific Computing Associate (SCA) position represents an alternative to traditional scientific roles (e.g. postdoc) and provides an ideal environment to establish a career in computational research or software engineering. The position aims at developing qualifications and experience in computational research and professional software engineering in a research environment that enables the candidate to pursue their future career in science or industry. The SCA position is a time-limited appointment for 24 months, with discretionary renewal for a final 12-month term (maximally 36 months in total.) We are seeking a talented and motivated computational scientist to develop and deploy cutting-edge experimental platforms that integrate dynamic virtual reality environments with precise neural and behavioral measurements in animal subjects (fish and flies) for the study of social learning and collective behaviors. This will require synchronization of and logging from many system components (video acquisition, animal tracking and pose estimation, microscopy image acquisition and/or physiological recordings, video game engines, multiple displays, etc.) as well as development of geometrically precise, reconfigurable, closed-loop virtual social paradigms that can be reproduced across animal subjects. By integrating real-time behavioral and neural measurements with virtual social environments, we enable neuroscientists to measure and model social behaviors in new and creative ways. You’ll work in close collaboration with Scientific Computing, MCN-NET, and the Schulze and Otopalik Labs. As part of a highly interdisciplinary and collaborative team of computational scientists, software and AI engineers, and neuroscientists, you’ll have access to high-performance workstations, CPU/GPU clusters, and experimental systems tailored for fish and fly research. This role will necessarily involve both software development and software-hardware integration, with potential opportunities to collect key initial datasets and contribute to publication(s) with the Schulze and Otopalik labs.

Requirements

A degree in computational sciences or equivalent (M.Sc. or Ph.D).
Experience with C# and Python programming (async/multithreading) languages.
Experience with real-time programming.
Experience in solving complex problems independently.
Good communication skills, comfortable working collaboratively in a team environment.

Nice To Haves

Experience with machine learning, big data, signal processing preferred.
Experience with the following will be extremely useful: messaging frameworks (e.g. sockets; currently UDP client used to communicate between Bonsai RX, animal tracking software, DAQs, and Unity), ROS-like systems, machine vision, GPU programming, shared memory, agent-based modeling, and/or game development.

Responsibilities

Develop Software Architecture for Fish & Fly Experimental Systems Clock synchronization across heterogeneous data streams, determined by hardware/software.
Build a robust synchronization layer to ensure <8 ms end-to-end latency for sub-frame accuracy in high-frequency behavioral streams.
Rigorous metadata and I/O logging to guarantee reproducible analysis pipelines.
Debugging driver/firmware bottlenecks in DAQs, GPUs, and cameras. (Interfaces with diverse hardware: cameras, DAQ boards, GPUs, head-mounted displays, lasers/scanners.)
Performance optimization for low-latency, precise, & dynamic virtual environments.
Develop reproducible open- and closed-loop virtual social environments and interactions using virtual fish and flies with hard-coded, dynamic, or agent-based movement rules (working in close collaboration with experimentalists).
Document and Generalize Software Modules for Widespread Use Organize software packages that generalize across fish and fly experimental setups with intuitive user interfaces that can be implemented within and beyond the Janelia Research Campus.
Compose documentation for reproducibility, clear metadata standards, and user-friendly interfaces (modular APIs and wrappers so experimentalists can use GUIs/editors instead of diving into C#/low-level code).
Strong Git-based version control workflows and containerization for reproducible deployments.
Potential opportunity for publication of this suite of tools in a methodological journal.