Candidate Profile — Candidate #2
Position: Adobe — Applied Scientist - Multimodal (San Jose, CA; San Francisco, CA; Seattle, WA)
Candidate Location: San Francisco Bay Area
Experience: 11.5 years experience
Relevant Experience & Education Highlights
Extensive expertise in generative AI, diffusion models, multimodal systems, and robust AI aligns strongly with advancing IP-aware guardrails and inference-time alignment for Adobe Firefly.
1. Research & Technical Depth
Possesses PhD-level background in generative models, fine-tuning, and multimodal AI essential for IP-aware generative modeling.
- Led foundation AI research innovations on large language models, diffusion models, and multimodality at a major enterprise SaaS company.
- Developed SAC^3 for hallucination detection and mitigation in black-box language models, enhancing reliability and trustworthiness (EMNLP 2023).
- Advanced robust generative flow models using normalizing flows for uncertainty estimation in computational imaging like FastMRI reconstruction.
- Published 50+ peer-reviewed papers including first-author works at NeurIPS, CVPR, EMNLP, AAAI, with invited reviews at ICLR, ICML.
2. Scientific & Systems Rigor
Demonstrates rigorous experimentation, failure mode analysis, and large-scale systems experience for optimizing multimodal pipelines.
- Led team on AI robustness, safety, probabilistic generative AI, and anomaly detection at a national laboratory.
- Developed scalable distributed deep learning frameworks with Horovod for near-linear scaling on supercomputers.
- Directed 7 projects as Principal Investigator totaling over $6.4 million in funding for AI-accelerated scientific discovery.
- Conducted uncertainty quantification, Bayesian optimization, and self-supervised methods for multimodal failure modes.
3. Vision-Language & Multimodal Reasoning
Brings hands-on work in multimodality, embeddings, and reasoning to enable semantic IP understanding and safety-aware inference.
- Drove research on LLM reliability, post-training, retrieval-augmented generation, agents, and multimodality at a major enterprise SaaS company.
- Proposed density-based self-supervised anomaly detection leveraging probabilistic generative models and active learning.
- Explored interactive multi-fidelity learning for language model adaptation (NeurIPS 2023) and gradient leakage auditing in federated learning (CVPR 2022).
4. Collaborative Approach
Exhibits tech leadership and cross-functional experience to architect pipelines and drive research-to-product translation.
- Served as Tech Lead and Senior Staff Research Scientist, responsible for technical directions and tech transfer to business products.
- Led efforts in AI for science on supercomputers, including large-scale optimization and distributed training.
- Taught courses on probabilistic methods and solid mechanics as teaching assistant at a top research university.
Requirements & Candidate Alignment
| Adobe Requirement | Candidate Qualification |
|---|---|
| Education: PhD or MS in Computer Science, Machine Learning, AI, or related field | PhD in Computational Science and Engineering and MS in Applied Mathematics and Statistics from a top research university |
| Experience: 5+ years of experience in applied ML or generative AI research (industry or academia) | 11.5 years across senior staff research scientist, tech lead, and national laboratory roles focused on generative AI and robust ML |
| Technical Background: Strong background in large-scale generative models (diffusion models, multimodal transformers, autoregressive systems) | Diffusion models and large language models expertise from robust generative flows, hallucination detection, and multimodality projects |
| Expertise: Deep experience with model fine-tuning, alignment strategies, and representation learning; Expertise in Vision-Language Models or multimodal foundation models | Fine-tuning and alignment via post-training LLM reliability work; Multimodal and vision-language models in RAG, agents, and computational imaging |
| Proficiency: Proficiency in Python and modern ML frameworks (e.g., PyTorch), with experience of training and deploying large models | Python and PyTorch proficiency; trained large models with Horovod for distributed deep learning on supercomputers |
| Publications: Research contributions in controllable generation, alignment, AI safety, or multimodal learning; Publications in leading conferences (CVPR, ICCV, NeurIPS, ICML, ICLR, SIGGRAPH) | 50+ peer-reviewed papers including NeurIPS, CVPR, EMNLP, AAAI; contributions to AI safety, hallucination detection, and multimodal robustness |
| Scientific & Systems Rigor: Strong experimental development and statistical evaluation skills; Experience analyzing complex failure modes in multimodal systems; Understanding of large-scale inference systems | Experimental rigor in uncertainty quantification, anomaly detection, and scalable optimization; analyzed multimodal failure modes in imaging and LLMs |
| Collaborative Approach: Experience working in cross-functional research-to-product environments | Tech Lead driving innovations to product impact; PI on multi-million-dollar projects with team leadership at national laboratory |
If you would like to discuss this candidate or other critical roles, here is a link to my calendar to schedule a call:
Schedule a CallContact:
Finding the signal in the noise since 2005