Kunal Saini

AI Researcher

Bellevue, Washington, United States8 yrs 2 mos experience
Most Likely To SwitchAI ML Practitioner

Key Highlights

  • Led ML infrastructure cost optimization, saving over $10M annually.
  • Developed architecture for content moderation using LLM models.
  • Designed AI platform for personalized insurance products.
Stackforce AI infers this person is a Machine Learning Engineer with expertise in E-commerce and Cloud Computing.

Contact

Skills

Core Skills

Machine LearningCloud ComputingData EngineeringSoftware DevelopmentWeb DevelopmentMobile Development

Other Skills

AI Platform DesignAWSAgentsAgile MethodologiesAlgorithmsAndroidAndroid DevelopmentApache KafkaBack-End Web DevelopmentBig DataCC++CSSChrome Extension DevelopmentComputer Vision

About

Experienced software developer with 7 years of experience in the e-commerce domain. Interested to work on AI and data driven projects, that involve Machine Learning/Deep Learning supported via cutting-edge technologies and frameworks and powered by cloud(AWS). I'm a high-spirited guy who is highly adaptable, a flexible professional who embraces teamwork but also enjoys working independently and loves networking. Great interest in Data Science and strong research professional with a Bachelor's degree focused on Computer Science from IIITD. #datadrivesbusiness

Experience

8 yrs 2 mos
Total Experience
2 yrs
Average Tenure
4 yrs
Current Experience

Amazon

SDE 2 (MLE)

Jun 2022Present · 4 yrs · Bellevue, Washington, United States · On-site

  • Org: Trust and Privacy at Amazon
  • Team: Alexa Sensitive Content Intelligence
  • 1. Lead organization-wide ML infrastructure cost optimization initiative, architecting solutions to reduce hardware costs from $30M to $20M (33% reduction) for 10K TPS content moderation system. Saving more than 10M $ annually.
  • 2. [Alexa+ Content Moderation] Leading and Developing the Architecture for the content moderation inference platform, that include hosting LLM models for offline response evaluation and (Building RAG infrastructure) embedding models with in-memory vector databases for prompt engineering that will enable LLM to generate moderated response in run time itself, along with context summarisation model. Along with that working on inference servers and libraries for latency & throughput optimizations (VLLM, DeepSpeed, Triton Inference, TensorTR, TGI).
  • 3. Leading the ML front and productionisation of BERT base Content Moderation Models serving as LLM guardrails.
  • (Model Deployments, ML Infrastructure, Scaling and Hardware Optimisations, Load & Integration Testing)
  • 4. Lead Multimedia project that enabled the team to explore new dimensions and deploy Computer Vision models for images/multimedia sensitivity detection and onboard new clients and use cases apart from existing NLP ones.
  • 5. Developing/Enhancing ML and Data platforms
  • 6. Creating Data Pipeline(ETL) for data annotations and ML model performance metrics generation, utilising spark functionalities
  • 7. Productionisation of text moderation(NLP) and image moderation(CV) ML/DL models
Machine LearningDeep LearningAWSComputer VisionData PipelineCloud Computing

Uber

SDE 2 (Data + ML)

Nov 2021Jun 2022 · 7 mos · Bangalore Urban, Karnataka, India

  • Team: Global Intelligence
  • 1. Productionalizing Uber's category position for a specific country
  • 2. Developing Monitoring and Data quality Checks pipelines
  • 3. Contributing to Dual Apper Users model productionalization
  • 4. Supporting, fixing, and improving data pipelines
  • 5. Developing rider-driver cancellation model as part of Uber wide Hackathon
Data Quality ChecksMonitoring PipelinesModel ProductionizationMachine LearningData Engineering

Expedia group

3 roles

Machine Learning Engineer II

Mar 2021Nov 2021 · 8 mos

  • Team: SPDE, Supply-Insights
  • 1. Productionalization of various Machine Learning and Deep Learning Algorithms.
  • 2. Building up ML and Data Pipelines.
  • 3. Supporting, improving, and migrating existing pipelines to a unified platform.
  • 4. AWS Sagemaker POC
  • Working with various tools/technologies:
  • Python, Scala, Spark, Airflow, Docker, Shell Scripting, Hive, Presto, AWS Services (Sagemaker, S3, EMR, EC2, Athena), Python libraries (scipy, sklearn, numpy, pandas), Deep learning Framework (Tensorflow/Keras), Qubole, Databricks, Notebooks, Hue
Machine LearningDeep LearningAWSData PipelinesCloud Computing

Software Dev Engineer II

Promoted

Mar 2020Mar 2021 · 1 yr

  • Team: Insurance
  • 1. Designing the AI platform for the team to empower various use cases; worked on Multi-Arm Bandit (Reinforcement Learning) based benefit segmentation use case in order to provide more personalized insurance products to the customer.
  • 2. Designing and Refactoring existing configuration management system to scale for multiple use cases and having version control supported via GIT Api's
  • 3. Driving ML Project from scratch i.e building model deployment service, data transformation/ feature selection, training and storing the result, logs for enhancement, and also collaborated with various Data Scientists across Geographies.
  • 4. Contributed towards developing a centralized email notification service for insurance, migrating from Mongo DB to Redis over Bull's infra. And making services loosely coupled.
  • 5. Integrated standalone insurance cancel skill with the ChatBot to save cost up to $1M per annum and load over our call centers.
AI Platform DesignReinforcement LearningModel DeploymentMachine LearningSoftware Development

Software Dev Engineer I

Jul 2018Mar 2020 · 1 yr 8 mos

  • Team: Insurance
  • 1.Developed plug and play; where UI is served via ReactJS technology, single click(payment) insurance module over the trip confirmation page to reduce friction over the checkout page.
  • 2. Enabled online insurance booking for customers who book through travel agents and helping in earning $1.5M per annum.
  • 3. Revamped post-purchase insurance; fixing bugs, improving logging and monitoring, adding retry logic, and fallback to the services and saving us $300K per annum.
  • 4. Developing various Test and Learns for the insurance module over the Checkout Page.
ReactJSUI DevelopmentInsurance Module DevelopmentWeb DevelopmentSoftware Development

Indraprastha institute of information technology, delhi

2 roles

Teaching Assistant

Jan 2018Apr 2018 · 3 mos · New Delhi Area, India

  • Course: Operating System

Teaching Assistant

Aug 2017Nov 2017 · 3 mos · New Delhi Area, India

  • Course: Image Analysis

Expedia, inc.

Intern

Jun 2017Jul 2017 · 1 mo · Gurgaon, India

  • Developed Chrome Extension for Flight Price Trend Analysis; used DynamoDB and consumed data via Kafka.
Chrome Extension DevelopmentData AnalysisSoftware Development

Houssup.com

Android Internship

Jun 2016Jul 2016 · 1 mo · Gurgaon, Haryana, India

  • Build Android application from scratch.
Android DevelopmentMobile Development

Education

Indraprastha Institute of Information Technology, Delhi

Bachelor's degree — Computer Science

Jan 2014Jan 2018

HANS RAJ MODEL SCHOOL

+2 — SCIENCE

Jan 2013Jan 2014

Stackforce found 100+ more professionals with Machine Learning & Cloud Computing

Explore similar profiles based on matching skills and experience