Research Scientist, Video & Multimodal

Job not on LinkedIn

🕒 September 24

🇺🇸 United States – Remote

💵 $160k - $185k / year

⏰ Full Time

🟡 Mid-level

🟠 Senior

🧬 Research Scientist

👻 Ghost score 20%

What it means: how likely this posting is sitting open without active hiring.How to read it: 60% or higher means we think this is a ghost. Lower scores are less of a worry.How we score it: age, reposts, open roles vs team size, how fast jobs close, and whether pay is listed.Fine print: an estimate from public data, not a judgment of the company. Real jobs can score high, so treat it as a hint, not a fact.

Apply Now

Find Similar Remote Jobs

Set Up Email Alert For New Jobs

📊 Check your resume score for this job

Improve your chances of getting an interview by checking your resume score before you apply.

Upload Resume

InnoData

WebsiteLinkedInAll Job Openings

2 - 10 employees

Founded 2019

🤝 B2B

💼 Consulting

🌍 Social Impact

B2B • Consulting • Social Impact

InnoData is INNOvation DATA SCS, an Italian social cooperative based in Foggia that identifies itself as a provider of technological solutions. The company website (currently under maintenance) highlights "Soluzioni tecnologiche" (technological solutions) and emphasizes social impact ("Impatto sociale"). Contact details listed include Via Francesco Crispi 65, 71121 Foggia, Italy. Based on the available information, InnoData appears to operate at the intersection of technology and social impact, likely offering tech-focused services to other organizations.

📋 Description

• Partner directly with customers and frontier labs building video understanding, video-language, and video-generation models • Define how video data is designed, structured, and evaluated for video and multimodal models • Translate model requirements into data specifications covering modalities, annotation schemas, sampling, and evaluation criteria • Build evaluation methodologies for video understanding, including temporal grounding, long-context reasoning, multi-turn, cross-modal, and retrieval-and-grounding evaluation • Build evaluation methodologies for video generation, including fidelity, temporal coherence, and physical plausibility • Decide how video is structured, enriched, and sampled, including messy, domain-specific footage • Run experiments and ablations connecting data choices to measurable model improvements • Design adversarial and stumping evaluations to identify system failures and improve data • Publish benchmarks, methodology, and papers • Work with annotation teams, subject-matter experts, and synthetic-data pipelines on collection and labeling plans

🎯 Requirements

• Roughly 5+ years of hands-on industry experience in video understanding or multimodal ML • A Bachelor's degree in computer science, electrical engineering, or a related technical or quantitative field is required • An advanced degree (MS or PhD) in a relevant field is preferred • Trained and evaluated video or multimodal models, with strong PyTorch fundamentals • Fluency in ffmpeg and decord pipelines, temporal and COCO-style annotation, WebDataset, Parquet and Arrow, and HuggingFace datasets • Experience fine-tuning large video or vision-language models with HuggingFace transformers, PEFT, and efficient inference • Experience with long-form video, streaming, temporal segmentation, or synthetic video generation • Experience building evaluation sets, calibrating difficulty, and defining quality video data for objectives • First-author publications or strong open-source contributions at venues such as CVPR, ICCV, ECCV, NeurIPS, or ICLR • Ability to work with research scientists and explain data and modeling decisions clearly to expert and non-expert audiences • Rigorous, reproducible approach to experiments and documentation • Interest or hands-on experience in responsible-AI evaluation and red-teaming is a bonus