Reetesh Mukul

AI Researcher

Bengaluru, Karnataka, India19 yrs 8 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Expert in developing large multi-modality models.
Strong background in computer vision and deep learning.
Proven track record in architectural development and model training.

Stackforce AI infers this person is a Computer Vision and Machine Learning expert with a focus on advertisement technology.

Contact

Skills

Core Skills

Computer VisionDeep LearningMachine LearningSoftware DevelopmentAndroid Development

Other Skills

AdSDKAdobe LightroomAlgorithmsArchitectural DevelopmentArchitectureArtificial Intelligence (AI)Bayesian ProgrammingBoost C++CC++CNNCognitive AnalysisComputational AdvertisementConvolutional Neural NetworkConvolutional Neural Networks (CNN)

About

I am currently working as a Senior Applied Principal Scientist in Computer Vision and Large Multi-Modality Models (LMM) domain. I specialize in developing large multi-modality models at various levels, including tasks specific to video and documents, as well as architectural modifications to incorporate multi-domain prompts and temporal aspects.

Experience

19 yrs 8 mos

Total Experience

2 yrs 10 mos

Average Tenure

2 yrs 6 mos

Current Experience

Oracle

Senior Principal Applied Scientist

Dec 2023 – Present · 2 yrs 6 mos · Bengaluru, Karnataka, India

I am working on Large Multi-Modality Models with a focus on Document AI, Video Temporal Aspects, and Grounding. My work involves Model Training and Architectural developments.

Large Multi-Modality ModelsDocument AIModel TrainingArchitectural DevelopmentComputer VisionDeep Learning

Huawei singapore

Senior Principal Research Engineer

Dec 2022 – Nov 2023 · 11 mos · Singapore · On-site

My role is to conduct and coordinate Research Initiatives associated with Ad Creatives -- understanding the content of the Advertisements and the cognitive interactions between various Ad Elements and users. The area thus involves Computational Advertisement, Creatives Psychology, Computer Vision, Cognitive analysis and Bayesian Programming. Most of my work requires developing Deep Learning Vision Models, Text Attention Models involving Advertisements. I work extensively on Neural Architectures, Image Attention, Transformers, Saliency and Memorability.

Computational AdvertisementDeep Learning Vision ModelsText Attention ModelsNeural ArchitecturesImage AttentionTransformers+4

Huawei technologies india

Principal ML Engineer

May 2021 – Nov 2022 · 1 yr 6 mos · India

> Ad Creative -- Deep Learning-Driven Advertisement Creation. I work on Automatic Advertisement Creation; and Creative Ranking using various compositional aspects involving Advertisements. All factors, like -- Industry, Product, Market, Users, Arts; -- are involved. The system involved learning the causal impact of each component on the overall performance. My work further involves several CNN/Transfomer based tools for Visual Analysis, Cognitive Analysis, Bayesian Programming for Reasoning, a few NLP and Sequential tools for learning the text impact, and Graphical Neural Network for understanding the connection among components.
> Saliency Prediction -- Network Development to predict Saliency
> Blur Prediction -- Network Development to predict Blur
> Memorability -- Network Development to preduct Memorability

Deep LearningCNNTransformerBayesian ProgrammingNLPGraphical Neural Network+2

Adobe

Senior Computer Scientist

Sep 2016 – Apr 2021 · 4 yrs 7 mos · Bangalore

Sensei: ML Model development and deployment
Visual Similarity Model Development
Panoptic Segmentation: Module development on Post Processing side.
Tensor Library development for a subset of PyTorch
Metric development for ML models.
Lightroom Performance: Performance enhancement for Lightroom Classic Import, Grid, and Library
Markov Decision Process-based Dynamic Algorithm for Import: Parallel Distributed Batch Allocation based on MDP.
Performance Database for Lightroom: Light-weight embedded Performance Database that Profiles Lua code automatically.
Common Table Expressions for folders, keywords. Facilitates very fast lightweight low latency queries.
Feature Prediction: Natural Language Model for Photographic Features. Predicts future features based on previous features.
Analytics for Lightroom.
ML Models for Churn analysis and Garbage Collection

ML Model DevelopmentVisual Similarity ModelPanoptic SegmentationTensor Library DevelopmentPerformance EnhancementMarkov Decision Process+2

Flipkart

SDE III

Aug 2015 – Aug 2016 · 1 yr · Bengaluru Area, India

Responsible for Architecture of AdSDK for FlipKart's Android App.. This is a highly scalable, concurrent architecture that joins avenues of AdNetwork and Publishers. Careful modelling of this system involved provisioning Parallel Functional Programming for Android along the lines of composable Moands. Currently participated in efforts for CTR prediction for Native Ads.

ArchitectureAdSDKParallel Functional ProgrammingSoftware DevelopmentAndroid Development