Shakti Kumar

AI Researcher

Toronto, Ontario, Canada8 yrs 3 mos experience
Most Likely To SwitchHighly Stable

Key Highlights

  • Designed ML quantization techniques reducing embedding costs by 8x.
  • Led Instagram Stories Ranking, driving $45M in revenue.
  • Expert in machine learning architectures and optimization.
Stackforce AI infers this person is a Machine Learning expert in AdTech and Social Media.

Contact

Skills

Core Skills

Machine Learning

Other Skills

Quantization TechniquesTransformerVariational AutoencodersPythonData StructuresGitC++CMatlab

About

I design core modelling architectures of ads recommenders at meta. Previously designed instagram stories/reels recsys. https://www.cs.toronto.edu/~shaktik/

Experience

8 yrs 3 mos
Total Experience
1 yr 8 mos
Average Tenure
3 yrs 11 mos
Current Experience

Meta

2 roles

Research Scientist

Promoted

Apr 2024Present · 2 yrs 1 mo

  • [Ads Recommendation] I research Ads RecSys architectures at Meta's Monetization org. In particular I design ML based quantization techniques which reduce cost of high dimensional embeddings by upto 8x to use in RecSys models at scale. See our work here: http://papers.adkdd.org/2025/papers/adkdd25-ramasamy-semantic.pdf
Machine LearningQuantization Techniques

Research Scientist

Jun 2022Mar 2024 · 1 yr 9 mos

  • [Instagram Ranking] I led the core modelling workstream at Instagram Stories Ranking driving key architectural improvements. I produced multiple launches with topline wins cumulating ~45M in overall IG revenue. I drove end-to-end processes: from developing neural architectures to leading launch processes. My ownership and XFN collaboration across Instagram’s ML teams optimized our GPU footprint significantly, led to substantial modelling+infra innovations and opened new investment directions.
Machine Learning

Vanguard

2 roles

Research ML Scientist

Jan 2021Jun 2022 · 1 yr 5 mos · Toronto, Ontario, Canada

  • Leading efforts on devising novel machine learning algorithms in Few Shot learning through statistics and neural networks.
  • Contributing to business use cases in multi-dimensional sentiment prediction/analysis of textual contents.

Research Intern

May 2020Dec 2020 · 7 mos · Toronto, Ontario, Canada

  • Researcher at the Center for Analytics and Insights, Vanguard working in unsupervised latent representation learning. Exploring Transformer based VAEs for improving disentangled latent representations.

University of toronto

2 roles

Teaching Assistant

Jan 2020May 2020 · 4 mos

  • Teaching Assistant for CSC384: Introduction to Artificial Intelligence at the Dept. of Computer Science supervised by Prof. Bahar Aameri and Prof. Sonya Allin.

Teaching Assistant

Sep 2019Dec 2019 · 3 mos

  • Teaching Assistant for CSC108: Introduction to Python at the Dept. of Computer Science supervised by Prof. Jennifer Campbell and Mario J. Badr

Cisco

2 roles

Software Developer

Jul 2018Aug 2019 · 1 yr 1 mo

  • Software Developer in Data Center Networks team developing Analytics and Learning models for Service Assurance in Cisco Data Centers globally.

Intern

Jan 2018Jun 2018 · 5 mos

  • Developed LSTM bots for Conversational Workplace AI Software using RASA and Cisco's open-sourced MindMeld libraries.

Sandisk®

Software Tools Intern

Jun 2017Jul 2017 · 1 mo · Bengaluru

  • CICD Team

Indian institute of technology, hyderabad

Research Fellow

May 2016Dec 2016 · 7 mos · Hyderābād Area, India

  • Networked Wireless Systems (NeWS) Lab https://newslab.iith.ac.in
  • Researcher in Access Barring Schemes for congestion control in MTC traffic in LTE for IoT applications

Education

University of Toronto

MSc. Applied Computing

Jan 2019Jan 2021

RV College Of Engineering

Bachelor’s Degree — Computer Science

Jan 2014Jan 2018

Stackforce found 100+ more professionals with Machine Learning

Explore similar profiles based on matching skills and experience

Shakti Kumar - AI Researcher | Stackforce