Senior AI Engineer (Driving VLM/VLA)
- AI
- Temporal
- Computer Vision
- Machine Learning
- PyTorch
About the Team & Mission
42dotμ Senior AI Engineer (Driving VLM/VLA)λ μλ°±λ§ κ±΄μ μ€μ μμ¨μ£Όν λ°μ΄ν°λ₯Ό νμ©ν΄ μ°¨μΈλ Driving Foundation Modelμ κ°λ°ν©λλ€. Vision-Language-Model(VLM), Vision-Language-Action(VLA), λ©ν°λͺ¨λ¬ AI κΈ°μ μ ν΅ν΄ μ°¨λμ΄ λ³΅μ‘ν λλ‘ νκ²½μ μ΄ν΄νκ³ νλ¨νλ©° νλν μ μλλ‘ λ§λλ ν΅μ¬ μν μ μνν©λλ€. λκ·λͺ¨ GPU μΈνλΌμ λ μμ μΈ μμ¨μ£Όν λ°μ΄ν°μ μ κΈ°λ°μΌλ‘ driving scene understanding, multimodal reasoning, VLM/VLAλ₯Ό ν΅ν΄ μμ¨μ£Όν μμ€ν μ μΈμ§, νλ¨, κ³ν μ±λ₯μ κ³ λνν©λλ€.
Responsibilities
μμ¨μ£Όν λ°μ΄ν°λ₯Ό νμ©ν Vision-Language-Action κΈ°λ° driving foundation model μ€κ³ λ° κ°λ°
Camera/video, map, trajectory, action, language λ± multimodal sequential data κΈ°λ° λͺ¨λΈ νμ΅ λ° νκ°
Driving scene understanding, temporal reasoning, agent interaction modeling, risk/event understanding λͺ¨λΈ κ°λ°
Scene captioning, visual question answering, auto-labeling, data mining, retrieval λ± VLM κΈ°λ° μμ© κΈ°λ₯ κ°λ°
λκ·λͺ¨ multimodal dataset ꡬμΆ, νμ΅ recipe, evaluation benchmark, ablation μ€ν μ€κ³
Simulation, Data, ML Platform μ‘°μ§κ³Ό νμ νμ¬ λͺ¨λΈμ μμ¨μ£Όν μμ€ν μ ν΅ν©
Qualifications
Computer Vision, Machine Learning, Robotics, Autonomous Driving, Multimodal AI κ΄λ ¨ 5λ μ΄μμ μ°κ΅¬/κ°λ° κ²½ν λλ μ΄μ μ€νλ μλ λλ Foundation Model, Multimodal Learning, Generative AI λΆμΌμμμ λ°μ΄λ μ°κ΅¬ μ±κ³Ό(proven track record)λ₯Ό 보μ νμ λΆ
PyTorch κΈ°λ° deep learning model κ°λ° λ° νμ΅ κ²½ν
Vision model, video model, VLM, VLA, world model, trajectory prediction, imitation learning μ€ νλ μ΄μμ λν κΉμ μ΄ν΄μ ꡬν κ²½ν
μ΄λ―Έμ§/λΉλμ€/sensor/trajectory/language/action λ± multimodal λλ sequential data μ²λ¦¬ κ²½ν
Transformer, diffusion, autoregressive model, representation learning μ€ νλ μ΄μμ λν μ΄ν΄
λ Όλ¬Έ κΈ°λ° μμ΄λμ΄λ₯Ό ꡬννκ³ λκ·λͺ¨ λ°μ΄ν°μμ νμ΅, νκ°, κ°μ ν κ²½ν
λ¬Έμ λ₯Ό λ 립μ μΌλ‘ μ μνκ³ , μ€ν μ€κ³λΆν° λͺ¨λΈ κ°μ κΉμ§ μ£Όλμ μΌλ‘ μνν μ μλ μλ
Preferred Qualifications
μμ¨μ£Όν λλ robotics foundation model κ°λ° κ²½ν
VLM/LLM fine-tuning, instruction tuning, multimodal alignment κ²½ν
VLA, behavior cloning, imitation learning, trajectory planning, policy learning κ²½ν
Action-conditioned video generation, world model, 4D scene modeling κ²½ν
BEV, occupancy, map, trajectory, agent interaction λ± driving-specific representation κ²½ν
Distributed training, large-scale multimodal data pipeline, data mining pipeline κ²½ν
Closed-loop simulation λλ scenario-based evaluationκ³Ό λͺ¨λΈμ μ°κ²°ν΄λ³Έ κ²½ν
Interview Process
μλ₯ μ ν
μ½λ© ν μ€νΈ
1μ°¨ λ©΄μ (νμ, 1μκ° λ΄μΈ)
2μ°¨ λ©΄μ (λλ©΄ νΉμ νμ, 3μκ° λ΄μΈ)
μ²μ° νμΒ·μ μ¬
Additional Information
μ ν μ μ°¨λ μΌμ λ° μ§ν μν©μ λ°λΌ μΌλΆ λ³κ²½λ μ μμΌλ©°, κ° μ ν κ²°κ³Όλ λ±λ‘νμ μ΄λ©μΌλ‘ κ°λ³ μλ΄λ립λλ€.
μ§μμ μ μΆ μ μ£Όλ―Όλ±λ‘λ²νΈ, κ°μ‘±κ΄κ³, νΌμΈ μ¬λΆ, μ°λ΄, μ¬μ§, μ 체쑰건, μΆμ μ§μ λ± μ±μ©μ μ°¨λ²μ μꡬ κΈμ§λ μ 보λ μ μΈ λΆνλ립λλ€.
μ§μμ μ μ μ€ μ€λ₯κ° λ°μνκ±°λ κΈ°ν λ¬Έμ μ¬νμ΄ μμ κ²½μ°, recruit@42dot.aiλ‘ λ¬Έμν΄ μ£ΌμκΈ° λ°λλλ€.
κ΅κ°λ³΄νλμμ λ° μ·¨μ λ³΄νΈ λμμλ κ΄κ³λ²λ Ήμ λ°λΌ μ°λν©λλ€.
μ₯μ μΈ κ³ μ© μ΄μ§ λ° μ§μ μ¬νλ²μ λ°λΌ μ₯μ μΈ λ±λ‘μ¦ μμ§μλ₯Ό μ°λν©λλ€.
42dotμ μλ’°νμ§ μμ μμΉνμ μ΄λ ₯μλ₯Ό λ°μ§ μμΌλ©°, μμ²νμ§ μμ μ΄λ ₯μμ λν΄ μμλ£λ₯Ό μ§λΆνμ§ μμ΅λλ€.
μ§μμ λ΄μ© μ€ νμ μ¬μ€μ΄ λ°κ²¬λ κ²½μ°, μ μ¬κ° μ·¨μλ μ μμ΅λλ€.
μΈν°λ·° νλ‘μΈμ€ μ’ λ£ ν μ§μμμ λμνμ ννμ‘°νκ° μ§νλ μ μμ΅λλ€.
3κ°μμ μμ΅κΈ°κ°μ΄ μ μ©λ μ μμ΅λλ€.
Senior AI Engineer (Driving VLM/VLA) Β· 42dot