Akanksha Gupta

Senior Software Engineer

Kirkland, Washington, United States7 yrs 3 mos experience
Highly Stable

Key Highlights

  • Expert in Machine Learning and Cloud Infrastructure.
  • Proven track record in developing real-time applications.
  • Strong background in data management and algorithm development.
Stackforce AI infers this person is a Fintech Software Engineer with expertise in Machine Learning and Cloud Infrastructure.

Contact

Skills

Core Skills

Machine LearningCloud InfrastructureSoftware DevelopmentTelemetryData VisualizationComputer VisionAlgorithm DevelopmentSoftware EngineeringSoftware TestingData AnalysisData EngineeringData Management

Other Skills

Cloud TPUNetwork OptimizationOpen Source ContributionsTelemetry DashboardsASP.NETC#Unit TestingElasticsearchAWSJavaCDK PipelineNeural NetworksNatural Language ProcessingAlgorithm GenerationData Pipelines

About

I am a Software Engineer, working as part of the ML@Scale universe of Google Cloud. As an infra SWE, I worked on new initiative Multislice that introduces horizontal scaling in Cloud TPUs for improving performance of large Machine Learning models. In my day-to-day work, I carve solutions to scale the ML models across TPU chips. I pursued Masters in Computer Science at Northeastern University, with a key interest in Software Design and Algorithms. I worked on various academic projects in the field of Computer Vision (Image transformations), Machine Learning (Generative Adversarial Networks and Autoencoders) and Natural Language Processing (localizing key steps in instructional videos by using Neural Networks and Natural Language models). Previously, I interned at Microsoft Whiteboard where my project enhanced telemetry and monitoring for mapping request failures with user authentication schemes and device types. During my stint as an intern at Amazon Alexa, I worked on building an end-to-end pipeline for data gathering, setting up the Java repository within AWS to start an Elasticsearch service inside a VPC, allowing HTTP access via Nginx, Cognito and Federate authentication, and displaying data on a Kibana dashboard for the internal teams to analyze common failures and their root causes to be able to guide the managed partners in a better way. Prior to that, I worked with Futures team at the D.E. Shaw Group as a Senior Software Engineer, where most of my work was related to generating algorithms and building software to facilitate trading at the firm. Specifically, my work involved enhancing simulation software, creating data pipelines, building the test framework for opening and closing the trading day, generating portfolio management user interface, to name a few projects.

Experience

Google

Senior Software Engineer

Feb 2022Apr 2025 · 3 yrs 2 mos · United States

  • Cloud ML Compute Services - Cloud TPU & GPU - ML@Scale
  • Working as part of the infra and ML team.
  • Launched the groundbreaking Multislice program on Cloud TPUs in 2023.
  • Brought network optimizations to Multislice for faster workloads and optimized internal stack.
  • Contributed to the open source repositories of jax-ml, ml-auto-solutions and xpk.
Cloud TPUMachine LearningNetwork OptimizationOpen Source ContributionsCloud Infrastructure

Northeastern university

2 roles

Graduate Teaching Assistant

Sep 2021Feb 2022 · 5 mos · Boston, Massachusetts, United States

  • Teaching Assistant for the course CS 5310: Computer Graphics under Prof. Mike Shah.
  • Duties include grading student assignments, holding labs and resolving students' queries related to the course.

Graduate Research Student

Sep 2021Feb 2022 · 5 mos · Boston, Massachusetts, United States

  • Working on a project to build an Android application that allows capturing an image and calculates the total value of coins identified in the image, under the guidance of Prof Ravi Sundaram.

Microsoft

Software Engineer Intern

Jun 2021Aug 2021 · 2 mos · United States

  • Whiteboard Services
  • Enhanced telemetry dashboards of Whiteboard microservices to expedite debugging of issues among a client base of 20 million MAU.
  • Added an ASP .NET middleware in C# to persist user agent and authentication type from HTTP request headers into the Correlation Context.
  • Recorded a new metric for personal Microsoft accounts (MSA) authentication failures and ensured 100% code coverage in unit tests.
  • Delivered an NPM Typescript project that displays components from a Figma design given its URL onto Whiteboard, as part of Hackathon.
  • Shadowed on-call for a week and carried out deployment across multiple regions. Actively participated in design meetings and code reviews.
Telemetry DashboardsASP.NETC#Unit TestingSoftware DevelopmentTelemetry

Amazon

Software Engineer Intern

Jan 2021Apr 2021 · 3 mos · Bellevue, Washington, United States

  • Alexa Smart Home Partner Outages API and Dashboard
  • Drafted a customizable and interactive dashboard with Elasticsearch for the Alexa analytics and dev teams to visualize customer outages.
  • Developed a CDK pipeline and Java lambda functions to gather data from DynamoDB and RedShift with a cross-account SNS/SQS setup.
  • Hosted AWS services inside Virtual Private Cloud and allowed HTTP access to Kibana inside the cloud by configuring Nginx proxy server.
  • Added HTTP authentication by leveraging Cognito and Federate, along with IAM roles and LDAP groups for managing user access.
  • Shadowed on-call for a week, resolved customer tickets and bugs.
ElasticsearchAWSJavaCDK PipelineCloud InfrastructureData Visualization

Northeastern university

3 roles

Graduate Researcher at MCADS Lab

Sep 2020Dec 2020 · 3 mos · Boston, Massachusetts, United States

  • Worked on a research project as part of the Mathematical, Computational and Applied Data Science Lab (http://khoury.northeastern.edu/home/eelhami/mcadsLab.htm) with Yuhan Shen under Prof Ehsan Elhamifar. Developed a Machine Learning algorithm to extract the grammar of complex tasks from instructional videos. Identified closely-coupled procedure segments by deploying a Neural Network on the word embeddings generated with Word2Vec, ELMo, Bert and GloVe.
Machine LearningNeural NetworksNatural Language ProcessingComputer Vision

Undergraduate Teaching Assistant

Sep 2020Nov 2020 · 2 mos · Boston, Massachusetts, United States

  • Teaching Assistant for the course CS 4500: Software Development under Prof. Matthias Felleisen and Prof. Jason Hemann.
  • Duties include grading student assignments, holding labs and resolving students' queries related to the course.

Graduate Teaching Assistant

May 2020Aug 2020 · 3 mos · Boston, Massachusetts, United States

  • Teaching Assistant for the course CS 5310: Computer Graphics under Prof. Mike Shah.
  • Duties include grading student assignments, holding labs and resolving students' queries related to the course.

D. e. shaw india private limited

3 roles

Senior Member Technical

Jan 2018Dec 2019 · 1 yr 11 mos

  • Worked in the Futures team that trades in Futures instrument around the globe.
  • Languages, libraries and Tools
  • Languages & OS: Python, Java, Swing, C, C++, Perl, HTML, CSS, JavaScript, PHP, CoffeeScript, SQL, MySQL, NoSQL, Unix, Windows
  • Frameworks & Libraries: NumPy, Pandas, Matplotlib, TensorFlow, Keras, React/Redux, ElasticSearch, Kibana, Highcharts, REST, GIT
  • Highlight Projects
  • Portfolio Management Tool
  • Built a real-time React application from scratch, visualizing volatility, positions, market prices and trade paths for each trading instrument.
  • Implemented Python data services with in-memory and on-disk caching using HDF5 files, and used Apache servers to handle requests.
  • Added custom JavaScript implementation for zoomable and clickable market tree-map, time-series line charts and sortable table grids.
  • Successfully replaced the obsolete static webpage framework with a real-time UI with a significant improvement in design and speed.
  • Test framework for the software that kick-starts and closes a trading day
  • Implemented Python test framework, for the Futures team, to test its trading software for the future trading days against firmwide changes.
  • Added test framework for pre and post trading processes, along with an integration with the mock trade setup for placing real-time trades.
  • Built a scalable software by spawning processes in multiple threads and running bulk of the software on distributed systems.
  • Simulation environment to inspect the addition of a new product in the trading universe
  • Established simulation framework for new instruments to calculate returns, resulting in a profit/loss curve of each of the new instruments.
  • Wrote file parsers to add and parse the instrument raw data and store in pickle files and h5 caches, using Pandas and NumPy libraries.
  • Integrated the new software with production processes and reduced runtime of the application by 30% by using OOPS rules in Python.

Member Technical

Promoted

Jul 2016Dec 2018 · 2 yrs 5 mos

  • Worked in the Futures team that trades in Futures instrument around the globe.
  • Languages, libraries and Tools
  • Python, Numpy, Pandas, Matplotlib, In-house databases, SQL, NoSQL, Perl, GIT, Cluster resource management, Unix, Windows
  • Highlight projects
  • Data Extraction and Management Tool
  • Wrote downloaders and parsers for the trade reports published intermittently at ICE website and price data at Bloomberg SFTP terminal.
  • Established regular cron jobs to check data availability on the web servers and schedule the downloads and ensured correctness of data.
  • Added custom data pipelines for each download and on-disk caching layer for quicker access, by storing the data in pickle and h5 files.
  • Python visualizations for forecast position and risk data
  • Built software for plotting marginal forecast positions, PNL cones, price fluctuations and trade paths using NumPy and Matplotlib libraries.
  • Stored the plots as static webpages, scheduled the run time based on data availability and reduced runtime of the application by 40%.
  • Reporting tool for nearing trading limits
  • Established a real-time app that compared firm’s position in an instrument against the newly imposed spot limits by exchange regulators.
  • Warned traders through software-generated mails when an instrument approached the spot limit date or the position neared its limit.
Algorithm GenerationData PipelinesSimulation SoftwareAlgorithm DevelopmentSoftware Engineering

Intern

Jan 2016Jun 2016 · 5 mos

  • Languages, libraries and Tools
  • Java, C++, Python, Numpy, In-house databases, NoSQL, ElasticSearch, Logstash, Kibana, GIT, Cluster resource management
  • As a full-time intern for 6 months, worked on two fully independent projects.
  • Java API for Server mode in Real time Time-series Database
  • Implemented secure and concurrent read and write connections to the real time time-series database server implemented in Java.
  • Built Java and C++ client interfaces and ensured fast, reliable connections with NIO sockets serving one connection in each server thread.
  • UI tool to monitor latency and the root cause of system failures
  • Developed a custom ELK instance for Feeds application. Implemented log watcher using Lucene query language in Logstash and Python.
  • Established a secure ElasticSearch cluster to store the data in documents. Built a real-time Kibana UI with user-customizable graphs.
  • Used Elastic Watcher to continually process the input data and warn the respective teams whenever a process failure is identified.
Data ExtractionData ManagementPythonData Engineering

Education

Northeastern University

MS — Computer Science

Jan 2020Jan 2021

Thapar Institute of Engineering & Technology

Bachelor of Engineering - BE — Computer Engineering

Jan 2012Jan 2016

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

Explore similar profiles based on matching skills and experience