Abhishek Yadav

Software Engineer

Davis, California, United States7 yrs 6 mos experience
Most Likely To SwitchHighly Stable

Key Highlights

  • Led cross-functional initiatives improving data infrastructure efficiency.
  • Achieved significant latency reductions in routing systems.
  • Designed algorithms for ML data placement published in OSDI'24.
Stackforce AI infers this person is a Backend-focused Engineer with expertise in Infrastructure and Data Systems.

Contact

LinkedIn

Skills

Core Skills

Distributed SystemsData InfrastructureMachine LearningData IngestionServerless ArchitectureFront-end DevelopmentUser Experience

Other Skills

C++PythonC#Azure Cosmos DBAzure FunctionsReact.jsRedux.jsDjangoData PlacementQuery RoutingApache ThriftPython (Programming Language)Data StructuresAlgorithms.NET Core

About

Staff Software Engineer at Meta leading the data placement and routing infrastructure powering Meta's exabyte-scale data warehouse. Over the years, I've built systems that route billions of workload requests daily across datacenter regions, co-locate ML training data with GPU capacity, and balance analytics compute workloads - delivering significant latency improvements and O($M) annualized infrastructure cost savings. My work on ML data placement has been published in the OSDI'24 paper on Meta's ML scheduling infrastructure. Previously worked on structured data ingestion using low-latency, serverless architecture (Azure Functions) and React-based UX components at Microsoft. Skilled in C++, distributed systems, data infrastructure, and building reliable systems at scale.

Experience

7 yrs 6 mos
Total Experience
3 yrs 9 mos
Average Tenure
4 yrs 3 mos
Current Experience

Meta

2 roles

Staff Software Engineer, Meta SuperIntelligence Lab (MSL)

Promoted

Feb 2026 – Present · 2 mos · Davis, California, United States · Remote

Distributed SystemsC++Data Infrastructure

Staff Software Engineer, AI & Data Infrastructure

Jan 2022 – Feb 2026 · 4 yrs 1 mo · Davis, California, United States · Remote

  • Tech Lead for Resource Balance pod within Tetris - Meta's infrastructure for Hive table placement and workload routing across the exabyte-scale data warehouse, serving data ingestion, replication, analytics, and ML training workloads. Over the 4 years, led cross-functional initiatives across 6+ teams and 10+ engineers to improve reliability, efficiency, and scalability of data warehouse systems.
  • Designed ML training data placement algorithms to co-locate training data with GPU capacity, improving fresh-data demand colocation from 75%→96%. This work is detailed in Section 3 (Slow-path Data Placement) of the OSDI'24 paper on Meta's ML scheduling infrastructure.
  • Led warehouse resource balancing for Spark and Presto across both shared and dedicated compute clusters - reducing job queuing times by 4.5x, eliminating capacity imbalance related SEVs, and cutting cross-region network demand by 42% through data placement and real-time load-aware routing improvements, driving O($M) annualized infrastructure savings.
  • Rebuilt the Global Tetris Router from Python to C++, achieving 10x latency reduction (800ms→80ms), 70% infrastructure savings, and 99.9% uptime - now deployed across ~20 regions on ~150 servers, handling 1.5B+ requests/day as the single routing authority for all warehouse workloads.
  • Designed a table colocation framework for interactive warehouse workloads, achieving 98% routing latency reduction (3s→50ms), 81% reduction in cross-region reads, and 49% fewer latency SLO violations - enabling stronger data locality guarantees across all Meta interactive warehouse surfaces.
  • References:
  • OSDI'24 MAST Publication - https://www.usenix.org/conference/osdi24/presentation/choudhury
  • AI Infra@Scale conference talk referencing the OSDI'24 paper in the context of LLM training infrastructure - https://youtu.be/ELIcy6flgQI?t=1212
  • Software Engineer (Jan 2022) → Senior Software Engineer (Aug 2023) → Staff Software Engineer (Aug 2025)
C#Azure Cosmos DBData IngestionServerless Architecture

Microsoft

3 roles

Software Engineer II

Apr 2021 – Jan 2022 · 9 mos

  • [Microsoft Sports]: The team is responsible for E2E Sports experiences on different Microsoft products (Bing, MSN, Windows, etc.).
  • Ingested structured data of various entities (leagues/teams/players/matches) for different sports from upstream data provider.
  • Scaled the underlying azure functions based distributed low-latency data and UX infrastructure to support numerous sport leagues.
  • The work included adding caching layer, splitting the sport apps in multiple App Service Plans, avoiding SNAT ports exhaustion, making cosmos DB queries cost and time efficient, etc.
  • https://medium.com/microsoftazure/leveraging-azure-to-build-low-latency-microsoft-sport-experiences-4d041af0fbec
React.jsRedux.jsFront-end DevelopmentUser Experience

Software Engineer

Promoted

Oct 2018 – Apr 2021 · 2 yrs 6 mos

  • Worked in different teams under Search, Ads, News and Edge org.
  • Microsoft Sports [Apr, 2020 - Apr, 2021] - The team is responsible for E2E Sports experiences on different Microsoft products (Bing, MSN, Windows, etc.).
  • Edge browser (Personalized Experiences) [Jan, 2020 - Mar, 2020] - Worked on various features of Collections for chromium-based Edge. Collections is a new take on Favorites/Bookmark by Microsoft for its new browser. Specifically, I have worked on extending the ability to add a comment/note to collection item, adding the color palette in note card for changing the background color (similar to sticky notes), and creating a first-run experience to educate first-time Collections users. These new features are available from Edge 84 stable release.
  • Microsoft News (MSN) [Jan, 2019 - Dec, 2019] - Built various front end experiences that get viewed a million times a day across Microsoft News pages using a React/Redux tech stack, with performance, reliability, and uptime being key metrics. Some of the projects that I have worked on include Spotlight, US Election 2020, Blended Enterprise Page for Edge.
  • Work Blogs:
  • https://blogs.msn.com/election-2020-explore/
  • https://www.msn.com/en-us/news/elections-2020/polls
  • https://www.msn.com/en-us/news/spotlight
  • https://blogs.windows.com/windowsexperience/2020/03/30/the-top-10-reasons-to-switch-to-the-new-microsoft-edge/
Django

Software Engineering Intern

May 2017 – Jul 2017 · 2 mos · Bangalore, India

  • Arrived at a mapping between the files changed in a pull request to a collection of test cases that might get affected, via Area Path, and used weighted probability to predict their risk of failing.

Education

Indian Institute of Technology, Guwahati

Bachelor of Technology (B.Tech.) — Computer Science and Engineering

Jan 2014 – Jan 2018

Stackforce found 100+ more professionals with Distributed Systems & Data Infrastructure

Explore similar profiles based on matching skills and experience

Jiayang Liu

Jiayang Liu

高级工程师

at 华为

Singapore, Singapore3 yrs 6 mos exp
Software ArchitectureAndroid Development
Karan Kumar

Karan Kumar

Software Engineer

at Upswing Financial Technologies

Bengaluru, India5 yrs 9 mos exp
Ashish Sinha

Ashish Sinha

Founder

at Stealth Startup

New York, United States13 yrs 8 mos exp
Riyaz Panjwani

Riyaz Panjwani

Senior Software Engineer

at Apple

Seattle, United States6 yrs 6 mos exp
VK

Vijayalekshmi Kesavan

Data Scientist

at Modulo Bio

Mountain View, United States11 yrs 2 mos exp
Causal InferenceRegression Analysis
Prateek Vaibhav

Prateek Vaibhav

Staff Software Engineer

at Uber

Bengaluru, India10 yrs 6 mos exp
Anthony N.

Anthony N.

Quant Trading Analyst

at DRW

United States1 yr 6 mos exp
Rishul A.

Rishul A.

Engineering Manager II

at Uber

Bengaluru, India10 yrs 8 mos exp
Data ArchitectureData Pipeline ManagementData PlatformDatabase Service DesignMLOps