Please complete the CAPTCHA to continue
@deepmind.com
✓
LinkedIn matched
A concise factual answer block for searchers comparing this professional profile.
Jason Lin is listed as GenAI @ DeepMind, Google[x], Apple, Lyft | Machine Learning & NLP | 0-1 Builder ^ Investor based in San Francisco Bay Area, United States. AeroLeads shows a work email signal at deepmind.com and a matched LinkedIn profile for Jason Lin.
Jason Lin previously worked as Research Scientist at Google Deepmind and Research Engineer, Vision and Language at Google Deepmind. Jason Lin holds Bs, Computer Science And Business Administration from University Of Southern California.
This section adds company-level context without repeating Jason Lin's masked contact details.
AeroLeads found 1 current-domain work email signal for Jason Lin. Compare company email patterns before reaching out.
0-1 engineer, 30x hackathon-winner turned AI PhD: Red-teaming LLMs (MIT), video transformers (Oxford), Visual Q&A for blind, 3D LiDAR perception (Lyft), InfoSec (Spotify), Arm-CUDA kernel (NVIDIA), agents (DeepMind).Awardee of National Math Cup and IEEE innovation prize. VC fellow at KP, a16z, Pear, GGV Capital, Fusion Fund; I previously built Faber, an AI-driven food discovery network. Invited speaker at: UN, Boston Dynamics, Udacity, Duke and Stanford. Blog: hackjlin.wordpress.comex-consulting and finance; I read vividly about history & governance.
Listed skills include C++, Public Speaking, Microsoft Office, Leadership, and 46 others.
A career timeline built from the work history available for this profile.
London, London, Gb
MoE + LLM pretraining for multi-step reasoning & web automationGemini efficiency: sparse attention in token routing, dynamic-res vision encodersAgent tool use; action world model, instruction tuning on LLM-data
London, London, Gb
* Research: video LLMs, self-supervised transformers/Perceiver/ViT* Safety: prompt tuning, model calibration, RLHF hallucination, image-text critic model* Incubation: Flamingo finetune API, industry-first Image Q&A & YouTube UX research* Interviewer of L4-6 candidates; patent inventor; scholar mentor AI4science
Cambridge, Ma, Us
Data-driven red teaming, RLHF for aligning adversarial/truthful LLMs (Stanford Existential Risk Initiative)
San Francisco, Ca, Us
Deep learning for LiDAR + Camera sensor fusion, data-parallel training & eval, Spark ETL. Liaised ScaleAI & WandB CEO partnerships, mixed-precision C++/GPU on-car deployment.* PyTorch, TensorFlow, sklearn, AWS SageMaker, SQL, Eigen, OpenCV, Kubernetes, docker, Flyte, Buildkite CI, seaborn, wandb
San Francisco, Ca, Us
Co-founded deep learning group at Self-Driving unit. From scratch: point cloud voxel pipelines, (20x speedup) anchor/FOV tuning, BEV-3D camera calibration, CNN architectures, frustum visualization. Invited to dropout -> convert full time.
USC T20 startup: Food discovery with computer vision. Led product, on-device ML & proprietary FoodRank algorithm. Beta tested with local businesses, managed team of 10 w/ weekly sprints (alum went onto Stanford, Google, Meta).
Mountain View, Ca, Us
Tech lead Brain & TensorFlow scientists to accelerate pose estimation w/ NVIDIA int8 depthwise conv. Debugged asymmetric Tensor core I/O, quantized matmul kernels & bottom-up keypoint decoder. TFLite for CUDA initiative featured at company-wide ML conference.
Mountain View, Ca, Us
Tensorflow.js toolkit for interpretability, efficient CNNs: https://oecd.ai/en/catalogue/tools/advisjsAdVisJS.com: world's first real-time adversarial attacks playground brought to the web browser.
Palo Alto, California, Us
Advisor trained in Stanford VC curriculum. DM if you're building something new!
Menlo Park, Ca, Us
Selected 42/1000+ (4%) to inaugural cohort of technologists at a16z Startup School.
Menlo Park, Ca, Us
Highly selective (7%) cohort of VC fellows for deep tech & hardware focus. Week-long "MBA".
Atlanta, Ga, Us
Machine Learning & Perception Lab: Deformable 3D object detection with geometric deep learning on RGB-D point clouds from synthetic data. CUDA accelerated 3D RoIAlign.
Atlanta, Ga, Us
(Accepted CVPR '18 Demo, arXiv:1806.05660) Using deep neural network visualization and interactive computer vision methods to interpret what different classification models 'see'.
Stockholm, Stockholm County, Se
Organized executive ML workshops and piloted Intrusion Detection System (IDS) for ATO over network activity. Ablated novel features for kmeans anomaly classification. Semi-supervised ensembles to probabilistic risk score. Advised by Ads Experience team, ingested geo-filtered logs from GCP datalake, BigQuery & Splunk.
Shenzhen, Guangdong, Cn
Researched state-of-the-art Computer Vision and benchmarked Faster RCNN vs. single-stage object detection algorithms for network efficiency, multiscale prediction with CUDA class-rebalancing focal loss on PASCAL VOC in Caffe. Deployment supports WeChat vision.
All-USC team awarded 5 prizes by Facebook, Amazon, Qualtrics, Stanford: health-tracker / VR debias app.3rd in world finals at Facebook HQ. 2-time winner at Stanford TreeHacks 2017, 2019.
Pasadena, Ca, Us
Future of work: Jointly explored next-gen AR/VR HCI input interfaces, ie. gesture, gyroscope, Leap motion, keyboard, Myo EEG & gaze pointer with JPL OpsLab Human Interfaces Group.
New York, Ny, Us
Influencer marketing: matching brands <-> YouTube/IG creators. Series A exited to Warner Bros.
Cupertino, California, Us
macOS Performance: Designed backtracking algorithm to instrument sources of context switch latency from aggregate trace logs for performance analysis. Modified XNU scheduler to uncover bottleneck processes from coalesced concurrent I/O, timer and kernel interrupts with custom logging. Identified Xcode breakpoint register non-determinism via assembly triage and debugging.
Los Angeles, California, Us
Course Producer and Grader for core CSCI-201: Principles of Software Development. Taught class of ~100 multi-threaded programming, mutex/semaphore and OS fundamentals.
Los Angeles, Ca, Us
Work study: Tech consultant troubleshooting university-wide stakeholders' IT customer service issues.
Quick answers generated from the profile data available on this page.
Jason Lin is listed as GenAI @ DeepMind, Google[x], Apple, Lyft | Machine Learning & NLP | 0-1 Builder ^ Investor.
AeroLeads has found 1 work email signal at @deepmind.com for Jason Lin.
Jason Lin is based in San Francisco Bay Area, United States.
Jason Lin has worked for Google Deepmind, Mit Computer Science And Artificial Intelligence Laboratory (Csail), Lyft Level 5, Faber Labs, and X, The Moonshot Factory.
You can use AeroLeads to view verified contact signals for Jason Lin, including work email, phone, and LinkedIn data when available.
Jason Lin holds Bs, Computer Science And Business Administration from University Of Southern California.
Jason Lin is listed with skills including C++, Public Speaking, Microsoft Office, Leadership, Javascript, Powerpoint, Java, and Html.
Search by job title, company, industry, location, and seniority. Export verified B2B contact data when you need it.
Start free trial Search contacts