Sandeep Krishnamurthy
Backend EngineerEngineering - kiro.dev
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Engineering - kiro.dev
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
-Passionate about making foundational model (FMs) training and inference - faster, cheaper, easier. -Currently focussed on building MosaicML (DataBricks) FM serving product line. Make FM run in production faster/cheaper/simpler. - Passionate about people - Ambitious, visionary, high velocity teams are magic.
International Institute of Information Technology-Bangalore
Master of Technology (MTech), Information Technology
January 1, 2012 – January 1, 2014
Dr. Ambedkar Institute of Technology
B.E, Computer Science & Engineering
N/A – Present
Vijaya Composite PU College
PUC, Physics, Chemistry, Mathematics, Biology
N/A – Present
Vijaya High School
SSLC
N/A – Present
Amazon Web Services (AWS)
Director Software Engineering
January 1, 2026 – Present
Santa Clara, California, United States · On-site
Databricks
Software Engineer/Director - Databricks MosaicAI Model Training
October 1, 2023 – December 1, 2025
San Francisco Bay Area · On-site
Amazon Web Services (AWS)
Senior Engineering Manager - Amazon SageMaker
March 1, 2020 – September 1, 2023
Santa Clara, California, United States · On-site
Engineering at Google Cloud AI
October 1, 2019 – March 1, 2020
Sunnyvale, CA
Amazon Web Services (AWS)
SDE 3
March 1, 2019 – September 1, 2019
Amazon Web Services (AWS)
SDE 2
November 1, 2016 – March 1, 2019
Amazon
SDE 2 Machine Learning
February 1, 2016 – November 1, 2016
Amazon
SDE1 Machine Learning
April 1, 2015 – February 1, 2016
Amazon
SDE 1
July 1, 2014 – March 1, 2015
Amazon
SDE-Intern
December 1, 2013 – June 1, 2014
IIIT Bangalore
Teaching Assistant
July 1, 2013 – September 1, 2013
Intelligent System to Predict Success of a Movie
August 1, 2013 – Present
A prediction system can be built to predict if a movie is going to be a box-office hit or flop. Such a system might consider various information from different sources like Twitter, Youtube, Google Search keys, etc. and come up with a set of key features and characteristics that can tell if a movie is going to be success with a certain probability.
Picturesque: Geo-fencing based Android App
August 1, 2013 – Present
The Android app lets a user to create a geo-fence around a desired location (radius of the fence will be a user input), and displays all the nearby interesting places on the user's phone, whenever he enters the geo-fence he had created earlier. This is done by polling the lat-long coordinates of the user. The same app can be extended to display hospitals/restuarants/fire stations etc. in an area, according to the need.
Picvik (IBM TGMC Project)
January 1, 2013 – Present
An online photo and video sharing websites where user can create own accounts, upload pictures and videos and even rate, comment and discuss. Features a wall which streams all public uploads by all users in real time and a thread-based discussion board where users can discuss media between themselves. Built using JSP, Struts2, Metro CSS.
Recommendation System For Movies
January 1, 2013 – Present
In this project we have built a recommendation system for movies. Generally in sites like NetFlix, there will be numerous movies and users. It is very much important for site of this kind to understand tastes of its users and recommend appropriate movies to its users, so that their users spend more time on their site. Each user may have varied interests and tastes. Users may rate some movies they watch. Also, all users may not rate all movies. Hence, now recommendation problem turns out to be matrix completion problem. We solved this problem using low rank approximation scheme. We have used Singular Value Thresholding technique to fill out missing entries in the rating matrix. Then based on the guessed ratings we can recommend movies for the users. This system is also very helpful in e-commerce sites like Amazon, Flipkart etc... to recommend appropriate items to their customers. We have been able to achieve approximately 70 to 80% of acceptable recommendations. Technologies Used: Python, NumPy, SciPy
Virtual Machine Affinity Management
January 1, 2013 – June 1, 2013
In this project, we extended Red Hat's open source solution Virtual Machine Manager(VMM), to handle affinity among virtual machines. One or more virtual machines running on the system may be affine(closely collaborating) with each other. If virtual machines are closely related then migrating, individual virtual machines, to other physical host and hence separating virtual machines can cause severe hindrance in performance. We built modules, on top of VMM, to handle these virtual machine affinity rules. Administrators would now be able to create and manage affinity rules among virtual machines. Also, migration of virtual machine is also controlled based on defined affinity rules. Technologies used: Python, Libvirt API, PyGTK was used to UI
Virtual File System
August 1, 2012 – Present
A virtual file system (VFS) is an abstraction layer on top of a concrete file system that provides the ability manage information about files and folders present in the hard disk. The purpose of a VFS is to allow client applications to access different types of concrete file systems in a uniform way. A VFS can, for example, be used to access local and network storage devices transparently without the client application noticing the difference.The project supports basic operation for file management.
Getting Started with Google Kubernetes Engine
Coursera
June 24, 2026 – Present
Deep Learning Specialization
Coursera
June 24, 2026 – Present
Microsoft Certified Technology Specialist - Application Development using .NET 3.5
Microsoft
June 24, 2026 – Present
Cultural Fit Analysis
The candidate's career trajectory shows a strong inclination towards innovation, building new products/teams from the ground up, and working with cutting-edge AI/ML technologies. Their involvement in open-source projects and contributions to foundational AI infrastructure (DLAMI, DLC, AWS-PyTorch) suggest a collaborative and community-oriented mindset. The emphasis on democratizing AI and making complex systems accessible aligns with a culture of impact and broad adoption. The diverse project portfolio, from web applications to recommendation systems and virtual machine management, indicates a broad technical curiosity, although the recent focus is heavily on AI/ML.
Soft Skills & Operational Fit
The candidate's experience descriptions highlight strong leadership, team building, strategic planning, and customer obsession. Their roles at Amazon and Databricks demonstrate an ability to operate at a high velocity, scale organizations, and drive significant business impact. The focus on 'cheaper, faster, simpler' solutions indicates a pragmatic and efficiency-driven approach, which aligns well with operational excellence.