Soumi Sarkar

Software Engineer

India8 yrs 1 mo experience
Most Likely To SwitchHighly Stable

Key Highlights

  • Expert in building AI-powered mobile applications.
  • Led critical initiatives at Microsoft with a focus on performance.
  • Strong background in machine learning and data analysis.
Stackforce AI infers this person is a SaaS-focused Software Engineer with expertise in mobile and AI technologies.

Contact

Skills

Core Skills

AndroidKotlinIos DevelopmentMachine Learning

Other Skills

AIALPAlgorithmsAndroid DevelopmentCC++Cascading Style Sheets (CSS)Data StructuresDesign PatternsFluxHaskellJavaKotlin CoroutinesLinuxMicrosoft Office

About

I am a Software Development Engineer II at Microsoft, currently building M365 Copilot Chat on Android, where I work on delivering resilient, high-quality AI-powered experiences at massive scale. I specialize in building reliable, user-focused systems across Android and the web, with experience spanning enterprise and consumer products used by millions. At Microsoft, I’ve led and contributed to critical initiatives such as cross-tenant list sharing in To Do, alerting and support infrastructure, and core modules in the New Planner app—always with a focus on performance, correctness, and long-term maintainability. My work often lives at the intersection of platform thinking and product impact: optimizing sync pipelines, improving observability, collaborating across teams, and translating ambiguous problem statements into clean, scalable solutions. I’m comfortable working across stacks (Kotlin, React, TypeScript) and across boundaries—engineering, product, and design.

Experience

8 yrs 1 mo
Total Experience
2 yrs 8 mos
Average Tenure
6 yrs 1 mo
Current Experience

Microsoft

2 roles

Software engineer 2

Promoted

Sep 2021Present · 4 yrs 9 mos · Hyderabad, Telangana, India

AndroidKotlinModel-view-viewmodel (MVVM)Kotlin Coroutinesjetpack compose

Software Engineer

May 2020Sep 2021 · 1 yr 4 mos · Hyderabad, Telangana, India

KotlinAndroid DevelopmentRxJavaModel-view-viewmodel (MVVM)Android

Comviva

Mobile Application Developer

Sep 2018May 2020 · 1 yr 8 mos · Bangalore

AndroidiOS DevelopmentSwift (Programming Language)

Penn state college of information sciences and technology

Summer Research Intern (SN Bose Scholar 2017)

May 2017Jul 2017 · 2 mos · State College, Pennsylvania, USA

  • Research on "Speech deterioration in Alzheimer's Disease". I collected speeches and literary articles of politicians, author and other celebrities who had Alzheimer's. I applied Machine Learning Algorithms to these texts to help the machine learn about the statistical trends in the speeches of these celebrities over the years of their life. The deterioration can be in the form of loss in memory resulting in stammering speeches. or increased used of fillers, or increased used of pronouns or some trend in the use of the various parts of speeches etc. Several features were extracted from the texts collected. These features were fed to the machine to build the model. This helped us in classifying the celebrities into Alzheimer's group and Non- Alzheimer's group.
  • Tool used- Weka.
Machine LearningWeka

University of calcutta

Mining Twitter for Airlines Consumer Sentiment Analysis

May 2016Jul 2016 · 2 mos · Kolkata Area, India

  • Extracted tweets on various Indian airlines from Twitter using R and mined the data so as to analyse the sentiment of the passengers which can be positive, negative or neutral. Also judged the profit or loss of a particular airline company during a given time span by plotting Histogram from the responses obtained from the extracted tweets. Then the analysis of sentiment of the passengers using entity level analysis was performed. Entities such as food, hospitality, timings, etc were analysed separately. Designed the entire tool for the above task using R.
R

Indian institute of technology, kharagpur

Equivalence Checking for Scheduling Verification Using Value Propagation and Path Extension Method

May 2016Jun 2016 · 1 mo · Kharagpur Area, India

  • Checked two Finite State Machine with Data Path(FSMD) for equivalence using two processes- Path extension and Value propagation. Value propagation was proved better than path extension. Certain algorithmic modifications were made in the value propagation method as it showed wrong results for some test cases. Studied Normalised Sums and implemented in Haskell Programming language.
Haskell

Nit durgapur

PSO Implementation of Satellite Scheduling Tasks

Jan 2016May 2016 · 4 mos · Durgapur Area, India

  • Scheduled the tasks uploaded by the ground control station to the satellite so as to maximise the profit and keeping in mind the constraints of energy, memory and duration available. There are several works that were done previously on off-board satellite scheduling tasks where the ground station would schedule the tasks and upload to the satellite without knowing the energy or memory constraints of the satellite. In this work the scheduling was done by the satellite itself keeping the constraints in mind using Particle Swarm Optimisation(PSO) algorithm and the code was written in python.
Python

Education

National Institute of Technology Durgapur

Bachelor’s Degree — Information Technology

Jan 2014Jan 2018

M P Birla Foundation H S School

ISC — Science

Jan 2012Jan 2014

Stackforce found 100+ more professionals with Android & Kotlin

Explore similar profiles based on matching skills and experience