Nikunj Agrawal

Software Engineer

Mountain View, California, United States10 yrs 9 mos experience
Highly StableAI Enabled

Key Highlights

  • Expert in Machine Learning and Backend Development
  • Proven track record in AdTech and Mobile Applications
  • Strong foundation in optimization algorithms and AI
Stackforce AI infers this person is a Machine Learning Engineer with expertise in AdTech and Mobile Applications.

Contact

Skills

Core Skills

Machine LearningBackend DevelopmentResearchArtificial IntelligenceMobile Development

Other Skills

AlgorithmsAndroid DevelopmentC (Programming Language)C++Convex OptimizationData StructuresDistributed LearningGradient Descent AlgorithmsHTMLHadoopJavaJavaScriptLaTeXLinuxNaive Bayes Classifier

Experience

Google

Software Engineer

Jul 2015Present · 10 yrs 8 mos · Mountain View

  • Generalist SWE, who likes to solve product problems, working across the backend stack, solving quality (ML) and scalable infrastructure (handle xx million QPS) problems.
  • Current: Private ML Infra (TLM)
  • Past (as SWE-II, SWE-III, Senior SWE, Staff SWE):
  • Automated video ads for Youtube ads.
  • Privacy Preserving Remarketing for Display ads: Re-imagining the future of ads personalisation in Google Display Ads with privacy preserving technologies. Also worked on remarketing quality and dynamic ads modeling.
  • Google Sheets Natural language intelligence: Launched a feature from scratch in Google Sheets to allow users to query + analyze sheets data in natural language.
C++PythonMachine LearningJavaLinuxBackend Development

Microsoft

Research Intern

May 2014Jul 2014 · 2 mos · Bengaluru Area, India

  • Functional Approximation based Distributed Learning Algorithm
  • supervised by Dr. Dhruv Mahajan, Applied Sciences.
  • Studied Gradient Descent Algorithms like TRON, SVRG, L-BFGS used for Convex Optimization
  • and their implementation in All-Reduce framework according to the Statistical Query Model.
  • Worked on a new Algorithm that approximates the Convex Objective Function, uses it for local optimization in different nodes and reduces the Communication Cost for Convex Optimization.
  • Comprehensive Evaluation of the Algorithm was conducted in which our algorithm outperformed all other State-of-the-Art algorithms on 5 out of the 6 High-Dimensional Datasets considered.
Gradient Descent AlgorithmsConvex OptimizationDistributed LearningMachine LearningResearch

Semusi

Summer Intern

May 2013Jul 2013 · 2 mos · Noida Area, India

  • Smart Applications (Mobile Applications using Artificial Intelligence)
  • Smart Dialer - Worked on a Naive Bayes Classifier which would take Call Logs of a person for training and later use it to suggest possible people owner might call at a particular time.
  • It yielded an accuracy of approximately 61.1 % using a corpus of old call logs.
  • Algorithm implemented as an application for Android based Smart Phones.
  • Smart SMS - Cloud-based Android Application which acts as an automated expense manager keeping track of the transactions made electronically by the user, by reading SMS in the phone using Natural Language Processing and mails the user a monthly Expense Sheet.
Naive Bayes ClassifierNatural Language ProcessingAndroid DevelopmentArtificial IntelligenceMobile Development

Education

Indian Institute of Technology, Kanpur

Bachelor’s Degree — Computer Science

Jan 2011Jan 2015

Bhatnagar International School

High School

Jan 2006Jan 2011

Stackforce found 100+ more professionals with Machine Learning & Backend Development

Explore similar profiles based on matching skills and experience