Sr. Path Planning Algorithm Engineer (Multi-Camera Vision Path Planning)
스트라드비젼직무 설명
This position is with STRADVISION Seoul Office
[About STRADVISION]
We Empower Everything To Perceive Intelligently
STRADVISION is a global deep-tech company specializing in autonomous driving and vision AI, making life safer and better through AI-based camera perception technology. Since its founding in 2014, the company has developed and commercialized core AI technologies that accurately perceive and interpret the world around us. In 2026, STRADVISION became the first autonomous driving software company to be listed on KOSDAQ. Today, approximately 250 colleagues work together across eight offices worldwide.
STRADVISION has secured contracts to supply its camera perception software, SVNet, to 14 global automakers for more than 50 vehicle models. SVNet has been deployed in over five million production vehicles. STRADVISION won the Gold Award for “Best-in-Class Software for Perception Systems” at the AutoSens Awards for two consecutive years, in 2021 and 2022.
The company is now expanding the reach of vision AI beyond autonomous driving into robotics and physical AI. Join us in shaping the future of mobility and physical AI.
[Our Story]
🔎 STRADVISION’s Core Technology
🔎 STRADVISION's Perks & Benefits
Latest News from STRADVISION
📌 STRADVISION Unveils Its SVDataFlow Pipeline, Focusing on Technical Advancement and Scalability
[Our Technology]
- In 2019, STRADVISION became the world’s first deep learning-based startup to achieve ASPICE CL2 certification. In 2020, the company received the AVT ACES Autonomous Vehicle Innovation Award.
- STRADVISION won the Gold Award for “Best-in-Class Software for Perception Systems” at the AutoSens Awards for two consecutive years, in 2021 and 2022.
- In August 2022, STRADVISION closed a KRW 107.6 billion Series C funding round, with Aptiv of the United States and ZF Group of Germany participating as strategic investors, reflecting global recognition of the company’s technology.
- STRADVISION holds 168 U.S. patents related to AI and automotive technology and continues to invest actively in research and development to further strengthen its technological capabilities.
- In the first half of 2026, the cumulative number of vehicles equipped with SVNet exceeded five million worldwide. STRADVISION continues to grow despite an economic slowdown and intensifying industry competition.
- In 2025, STRADVISION ranked 10th in Korea’s AI competitiveness rankings, behind companies such as Samsung Electronics, NAVER, and LG.
[Mission of the Role]
The Role as "Architect of Driving Intelligence"
In a team pursuing a dual paradigm, the modern Senior Path Planning Engineer is elevated from a mere algorithm implementer to an "Architect of Driving Intelligence." They must be the bridge between the deterministic, interpretable world of classical robotics and the probabilistic, emergent world of deep learning. They analyze the failure cases of the modular planner (e.g., "Why did the A* planner freeze at this intersection?") to inform the design of a better reward function for the RL agent ("Let's add a penalty for prolonged hesitation"). Conversely, they observe the emergent, human-like behaviors of the E2E model and codify those heuristics as rules in the modular stack's behavior layer, enabling faster, near-term product improvements.
This role requires a systems thinker who designs the symbiotic relationship between the two approaches.
The Opportunity: Architecting the Vehicle's Brain
This role is not just about improving existing planners; it's about designing the next generation of driving intelligence. You will be responsible for the entire decision-making architecture of the vehicle, from navigating complex urban intersections to executing high-speed highway lane changes. You will drive innovation at the intersection of classical robotics algorithms and cutting-edge deep learning techniques.
This role is a unique opportunity to work on high-impact, cutting-edge research that directly contributes to the development of next-generation autonomous driving systems.
[Key Responsibilities]
Modular Stack (The Foundation)
- Combine traditional search-based (A*, D*), sampling-based (RRT*, BIT*), and optimization-based techniques with deep learning to maximize the efficiency and quality of path planning. For example, use deep learning-trained heuristics or cost maps to accelerate search speed, or generate reinforcement learning-optimized sampling distributions to quickly find better paths in complex environments.
- Develop and own the behavior planning layer, building robust state machines and decision-making logic for complex scenarios like unprotected left turns, merges, and obstacle avoidance.
- Design and validate interfaces that fully leverage the performance of the perception module by directly integrating probabilistic outputs and uncertainty estimates from the Perception team's deep learning-based models (e.g., object detection, semantic segmentation). Utilize the results of deep learning models that predict the behavior of surrounding objects to make proactive and safe decisions.
- Analyze real-world driving and simulation log data to identify failure modes, determine root causes, and drive continuous improvement of the planning stack.
End-to-End Stack (The Frontier)
- Lead the research and development of learning-based driving policies using Imitation Learning (IL) and Reinforcement Learning (RL)
- Design and formulate sophisticated reward functions for reinforcement learning that capture the nuances of safe, efficient, and comfortable driving, going beyond simple goal achievement.
- Research and implement hybrid architectures that combine the strengths of traditional planners and learned policies, such as using a classical planner to bootstrap an RL agent's initial learning or provide a safe fallback path.
- Collaborate with the Vision/DL engineers to co-design the architecture of the policy network, ensuring that the latent space representations are suitable for complex planning tasks.
Collaboration
- Collaborate with cross-functional teams, including machine learning engineers, software integration engineers, hardware platform engineers, and quality assurance, to integrate multi-vision E2E algorithms into ADAS systems.
- Participate in code reviews and knowledge-sharing sessions to foster a collaborative work environment.
Mentoring and Technical Guidance
- Mentor and provide technical guidance to junior/entry engineers.
[Basic Qualifications]
- Master's or Ph.D. in Computer Science, Robotics, Engineering, or a related field.
- 7+ years of hands-on development experience in path planning, behavior planning, or decision-making for autonomous robots or self-driving cars.
- Expert-level proficiency in C++ and strong Python skills for prototyping and data analysis.
- Deep theoretical understanding of robotics algorithms, including kinematics, dynamics, and probabilistic robotics.
- Proven experience leading complex software systems from conception to deployment.
[Preferred Qualifications]
- Publication record in top-tier robotics or AI conferences such as ICRA, IROS, CoRL, RSS, NeurIPS, or CVPR.
- Practical experience with one or more major E2E driving paradigms, such as imitation learning or reinforcement learning, in a robotics context.
- Experience with optimization-based trajectory generation and the use of numerical optimization solvers.
- Proficiency with ROS/ROS2, middleware (DDS), and modern software development practices (CI/CD, Bazel, Git).
[Application]
- Required: Resume / Thesis (for those who have a Master’s degree or above.)
- Optional: Cover Letter, Research/Project Portfolio (including publications, open-source projects, or patents), Other theses
- Application Review > Screening Test > Interview(s) > Reference Check > Offer > Onboarding
- Please be aware of that the recruitment processes & schedules may be changed depending on the job and/or other circumstances.
[Others]
- This position is open for ongoing recruitment and may close earlier once the role has been successfully filled.
- A 3-month probationary period will apply after joining as a regular employee.
- Visa sponsorship can be supported for eligible international candidates.
- If any false or inaccurate information is identified during the hiring process, the recruitment may be discontinued or an offer may be withdrawn.
STRADVISION stands for an open and respectful corporate culture because we believe the diversity helps us to find new perspectives.
STRADVISION ensures that all our members have equal opportunities –regardless of age, ethnic origin and nationality, gender and gender identity, physical and mental abilities, religion and belief, sexual orientation, and social background. We always ensure diversity right from the recruitment stage and therefore make hiring decisions based on candidate’s actual competencies, qualifications, and business needs at the point of the time.
Please feel free to contact us via our HR team e-mail if you have any questions.
[STRADVISION HR Team e-mail: [email protected]]
이 직무에 관심이 있으신가요?