{"data":{"id":106918,"slug":"ml-inzenyri","type":"job","title":"ML Engineers","description":"Dusejovsky Kamil, e-mail: info@revelrobotics.com\n\nWe are looking for an experienced professional who will lead, scale, and advance our training pipeline. Better models, larger datasets, more robust production stack, every percentage point of quality in strength, intent, and policy.\nWhat you will do\n\nLead the training pipeline at scale. Deploy improved skill models to real robots on a weekly cadence.\nAdvance model quality. Architecture, loss function design, tokenization, data curation, augmentation, evaluation. Own ablations.\nOwn training infrastructure. Multi-GPU, distributed runs, checkpointing, experiment tracking, automated evaluation.\nApproach data as a first-class concern. Curation, labeling, balancing, synthetic augmentation, sim-to-real.\nBe in the lab. Sit with the hardware and capture team. Close the loop between model and robot.\n\nWho we are looking for\n\nExperience in computer science, electrical engineering, robotics, ML, or a related field. Or completed work that demonstrates the same depth.\nNative work in PyTorch. Comfort with CUDA, multi-GPU, distributed training, mixed precision.\nHands-on experience with at least one of: imitation learning, behavior cloning, diffusion policies, vision-language-action models, RL on robots, sim-to-real.\nFluency across the modern robotics learning stack: large multimodal foundation models, photorealistic simulators, ROS-class middleware, modern policy architectures.\nStrong foundations in signal processing, sensor fusion, and high-dimensional time series modeling. Multi-rate synchronized sensor data doesn't intimidate you.\nYou take data seriously. Quality, balance, and curation are your job, not someone else's.","language":"cs","is_translated":true,"title_original":"ML in\u017een\u00fd\u0159i","description_original":"Dusejovsky Kamil, e-mail: info@revelrobotics.com\n\nHled\u00e1me zku\u0161en\u00e9ho pracovn\u00edka, kter\u00fd bude \u0159\u00eddit, \u0161k\u00e1lovat a posouvat n\u00e1\u0161 tr\u00e9ninkov\u00fd pipeline. Lep\u0161\u00ed modely, v\u011bt\u0161\u00ed datov\u00e9 sady, robustn\u011bj\u0161\u00ed produk\u010dn\u00ed stack, ka\u017ed\u00e9 procento kvality v oblasti s\u00edly, z\u00e1m\u011bru a politiky.\nCo bude\u0161 d\u011blat\n\n\u0158\u00eddit tr\u00e9ninkov\u00fd pipeline ve velk\u00e9m m\u011b\u0159\u00edtku. Nasazovat vylep\u0161en\u00e9 modely dovednost\u00ed na re\u00e1ln\u00e9 roboty v t\u00fddenn\u00edm rytmu.\nPosouvat kvalitu model\u016f. Architektura, n\u00e1vrh ztr\u00e1tov\u00e9 funkce, tokenizace, kur\u00e1torstv\u00ed dat, augmentace, evaluace. Vlastnit ablace.\nVlastnit tr\u00e9ninkovou infrastrukturu. Multi-GPU, distribuovan\u00e9 b\u011bhy, checkpointing, sledov\u00e1n\u00ed experiment\u016f, automatizovan\u00e1 evaluace.\nP\u0159istupovat k dat\u016fm jako k prvo\u0159ad\u00e9 z\u00e1le\u017eitosti. Kur\u00e1torstv\u00ed, labelov\u00e1n\u00ed, vyva\u017eov\u00e1n\u00ed, syntetick\u00e1 augmentace, sim-to-real.\nB\u00fdt v laborato\u0159i. Sed\u011bt s hardwarov\u00fdm a capture t\u00fdmem. Uzav\u00edrat smy\u010dku mezi modelem a robotem.\n\nKoho hled\u00e1me\n\npraxe v oboru informatika, elektrotechnika, robotika, ML nebo p\u0159\u00edbuzn\u00e9m oboru. Nebo realizovanou pr\u00e1ci, kter\u00e1 prokazuje stejnou hloubku.\nNativn\u00ed pr\u00e1ce v PyTorch. Pohodl\u00ed s CUDA, multi-GPU, distribuovan\u00fdm tr\u00e9ninkem, mixed precision.\nPraktick\u00e1 zku\u0161enost alespo\u0148 s jedn\u00edm z: imita\u010dn\u00ed u\u010den\u00ed, behavior cloning, diffusion policies, vision-language-action modely, RL na robotech, sim-to-real.\nPlynulost nap\u0159\u00ed\u010d modern\u00edm stackem robotick\u00e9ho u\u010den\u00ed: velk\u00e9 multimod\u00e1ln\u00ed foundation modely, fotorealistick\u00e9 simul\u00e1tory, middleware t\u0159\u00eddy ROS, modern\u00ed architektury politik.\nSiln\u00e9 z\u00e1klady ve zpracov\u00e1n\u00ed sign\u00e1l\u016f, senzorov\u00e9 f\u00fazi a modelov\u00e1n\u00ed vysokodimenzion\u00e1ln\u00edch \u010dasov\u00fdch \u0159ad. Multi-rate synchronizovan\u00e1 senzorov\u00e1 data t\u011b ned\u011bs\u00ed.\nBere\u0161 data v\u00e1\u017en\u011b. Kvalita, vyv\u00e1\u017eenost a kur\u00e1torstv\u00ed jsou tvoj\u00ed prac\u00ed, ne prac\u00ed n\u011bkoho jin\u00e9ho.","price":"60000.00","currency":"CZK","status":"active","noindex":true,"location":{"address":null,"full_address":null,"city":"Praha","country":"CZ","latitude":50.1029026,"longitude":14.3946353},"metadata":{"shift":"Pru\u017en\u00e1 pracovn\u00ed doba","region":"Hlavn\u00ed m\u011bsto Praha","employer":"REVEL Robotics s.r.o.","positions":5,"salary_to":200000,"profession":"ML in\u017een\u00fd\u0159i","start_date":"2026-08-01","salary_from":60000,"salary_unit":"K\u010d\/m\u011bs\u00edc","employer_ico":"29643465","suitable_for":["Vhodn\u00e9 pro cizince mimo EU","Zam\u011bstnaneck\u00e1 karta"],"min_education":"V\u0160 bakal\u00e1\u0159sk\u00e9","contact_person":"Kamil Dusejovsky","hours_per_week":40,"employment_type":"full_time","employment_types":["Pracovn\u00ed pom\u011br \u2013 pln\u00fd \u00favazek"],"reference_number":"33403200719"},"user_id":null,"is_sponsored":false,"views_count":0,"ai_views_count":17,"bot_views_count":1,"ai_vendor_counts":{"meta":5,"other":2,"openai":5,"anthropic":5},"visibility":"public","submission_source":null,"submission_ai_name":null,"has_owner_email":true,"can_contact_owner":true,"phone":null,"owner_email":"info@revelrobotics.com","images":[],"published_at":"2026-06-27T00:00:00+00:00","expires_at":null,"created_at":"2026-07-05T02:00:23+00:00","updated_at":"2026-09-20T02:24:25+00:00"}}