0
Publications
0
Followers
0
Following
2
Questions
About
Abhishek Das is a highly accomplished software engineer with over a decade of experience in building scalable and innovative technology solutions across multiple industries. Currently, he serves as a Software Engineering Lead at Salesforce, where he plays a crucial role in developing the Einstein Machine Learning platform for real-time inference, enabling enterprise-level AI solutions. His expertise in cloud computing, distributed systems, and artificial intelligence has significantly contributed to the success of Salesforce’s AI initiatives. Prior to this role, Abhishek worked at Microsoft for nearly five years, where he held key leadership positions and contributed to major projects such as Azure Dedicated and Azure Stream Analytics. At Azure Dedicated, Abhishek led the network control plane for building Baremetal as a Service for dedicated workloads, successfully onboarding Nutanix as the first workload. He made several improvements in the stack and mentored new engineers, helping to create a more efficient and high-performing team. His earlier work at Azure Stream Analytics involved leading the control plane effort for stream analytics on edge devices and delivering new features such as auto-scaling, parquet input, and customer-provided storage. Abhishek also played a critical role in enabling data plane migration from HDInsight to Service Fabric, ensuring a seamless and resilient orchestration process. His technical leadership and problem-solving skills were instrumental in the success of these projects.
Before joining Microsoft, Abhishek worked at Groupon as part of the computational marketing team, focusing on display ads engineering. He contributed to optimizing marketing solutions using data-driven approaches to improve customer engagement and business outcomes. His time at Dell EMC further refined his expertise in data protection and availability solutions. During his tenure, he received the prestigious EMC Excellence Award for providing critical patches that unblocked issues in a large CSV environment. He also holds a patent for optimizing the backup of clusters with multiple proxy servers, demonstrating his ability to innovate and solve complex technical challenges.
Abhishek’s academic journey began at Odisha University of Technology and Research, where he earned his Bachelor of Technology degree in Computer Science. He later pursued a master’s degree in Computer Science from Texas A&M University, where he was involved in several innovative projects. One of his notable projects was a distributed password cracker, which simulated a distributed environment with multiple worker nodes and a metadata server. This project allowed him to explore various aspects of distributed systems such as load balancing and fault tolerance. Another significant project was Trivia Pitcher, an online multi-user trivia system that created trivia questions from a Neo4j graph database built from Wikipedia’s XML dump. The system provided questions based on user difficulty levels, showcasing his interest in building intelligent and interactive systems.
Abhishek’s passion for technology extends beyond his professional work. He is dedicated to mentoring junior engineers and fostering a collaborative work culture. Known for his ability to think outside the box and deliver end-to-end solutions, Abhishek is widely respected in the tech community. His skills in distributed systems, cloud architecture, and AI have been endorsed by many colleagues and industry peers. With a relentless drive for innovation and a deep understanding of emerging technologies, Abhishek continues to push the boundaries of what is possible in software engineering.
Skills & Expertise
cloud computing
SPARK SQL
HIVE
Postgres SQL
Teradata
SQL
JavaScript
node.js
Artificial Intelligence
Distributed Systems
Java
Hadoop
Load Balancing
Auto Scaling
Emerging technologies
Redis
Edge Computing
Junit
Data Protection
Data Migration
cloud architecture
Real-Time Inference
Network Control
Baremetal Services
Service Fabric
Control Plane
Computational Marketing
Display Ads
Fault Tolerance
Neo4j Database
Innovation & Mentorship
test NG
Research Interests
cloud computing
Machine Learning
Software Engineering
Team Building
Computing
Mentorship Skills
Network Control Plane
Data Migration
Auto-Scaling Solutions
Service Fabric Orchestration
Display Ads Engineering
Fault Tolerance
Graph Databases (Neo4j)
Intelligent Systems
Hadoop
Distributed Systems
Connect With Me
Experience
Software Engineering Lead
- • Led design and development of a sophisticated multi-step multi-tenant Directed Acyclic Graph (DAG) execution service which enabled stitching of multiple steps required for successful inferences at extremely high scale with millisecond latencies. • Led Successful onboarding of several predictive and generative use cases with 100+ custom plugins written by app developers across multiple organizations. • Engineered on-demand scoring ML platform with a high-scale, low-latency multi-step DAG execution system for B2C applications. • Led the cross organizational effort of design and development of an innovative platform tailored for hosting Large Language Models (LLMs), specifically addressing CodeGen and Flow GPTs (Patent Pending). Ensured seamless handling of inference requests with an emphasis on achieving low latency and high scalability while leveraging some of current platform services. • Designed an innovative GDPR compliance violation resolution service capable of scanning petabytes of data. (Java, AWS EMR, Kubernetes, GRPC, REST, PyTorch, Python, Sagemaker, Inference Frameworks)
Senior Software Engineer Tech Lead
- 1. Azure Dedicated (10/20 - 9/21) - Tech lead (Network control plane) for building Baremetal as a service for dedicated workloads onboarding Nutanix as the 1st workload. - Made several improvements in the stack and processes around the team. - Mentored new engineers in the team. 2. Azure Stream Analytics (1/17 - 9/20) - Part of the core leads team to deliver Azure Stream Analytics dedicated Clusters (https://docs.microsoft.com/en-us/azure/stream-analytics/cluster-overview) - Lead control plane effort for Stream Analytics on edge devices (https://docs.microsoft.com/en-us/azure/stream-analytics/stream-analytics-edge) - Billing lead for Stream Analytics offerings. - Owned and designed the orchestrator on control plane responsible for managing the Stream Analytics jobs. - Delivered several new features like parquet input, auto scale ( https://docs.microsoft.com/en-us/azure/stream-analytics/stream-analytics-autoscale) customer provided storage ( https://docs.microsoft.com/en-us/azure/stream-analytics/data-protection) etc. - Enabled data plane migration from HDI to service fabric by moving some of migration resiliency and orchestration to control plane. - Mentored Junior engineers
Software Development Engineer
- • Developed tools and features with REST APIs for internal teams, utilizing machine learning algorithms for display ads bidding prices. • Led migration of the core bidding engine from Teradata to HDFS, optimizing performance. • Implemented a Real-Time Attribution data pipeline with a supporting dashboard for pricing updates. • Collaborated with marketing and product teams for new features and requirements. (Java, JavaScript, node, SQL, Postgres, Redis, Teradata, test NG, JUnit, Hadoop, Hive, Spark)
Software Engineer
- • Designed and implemented a high-performance Hyper-V Multi-Proxy backup architecture for Avamar and Networker's next release. • Enhanced backup efficiency over CSV and SMB storage, achieving significant performance improvements. • Secured a patent (9639428) for optimizing backup processes in clusters with multiple proxy servers. • Balanced backup loads efficiently among proxies, enabling parallel backups and boosting overall performance. • Provided engineering support and hotfixes for high-priority customer escalations, ensuring rapid issue resolution. (C++, C#, Win APIs, PowerShell, Socket Programming, IPC, XML)
Co-op, R&D IT
Web Developer(Student)
Education
Texas A&M University
Odisha University of Technology and Research (OUTR)
Projects
Improved HDFS Block Allocation Scheme
Modified the Block Allocation Scheme to keep related files together based on a tag associated with the file.
Dynamic Replication of Blocks in Hadoop
Hadoop HDFS makes 3 copies of each chunk by default. The project was to modify existing hadoop codebase to introduce dynamic replication for frequently requested blocks to improve response time.
Hadoop Block Placement
Designed an improved placement algorithm for HDFS by placing related files in the same subset of nodes.
Trivia Pitcher
An online multi-user trivia system which traverses a neo4j graph database created from Wikipedia xml dump to create trivia questions and provides questions based on the user’s difficulty level which is determined by several users and page attributes. (PHP, neo4j API)
Distributed Password Cracker
It creates a distributed environment with a single meta data server and several workers. We tried to implement few aspects of a distributed system such as load balancing, worker failure scenarios etc.
Certificates & Licenses (3)
Blockchain Basics
Blockchain technology demands attention. And it's not just due to the way it revolutionized monetary transactions. The concept of transactional trust through code has disrupted a variety of industries, offering ways to streamline everything from financial services to identity management. In this introductory-level course, learn what blockchain is and what it might mean to you.
Instructor Jonathan Reichental dives into the topic from a conceptual perspective, ensuring that a nontechnical audience can glean value from each lesson. Jonathan describes some of the current challenges with the internet, including existing risks and security problems. He goes over how traditional online databases function, so that you have a basis for how the blockchain redesigns this function. Plus, learn how the blockchain has become a potential solution for many of the existing limitations of online databases, which challenges must be solved in order to achieve broader adoption, and more.
https://www.linkedin.com/learning/blockchain-basics-14414119?trk=lynda_redirect_learning
Cloud Architecture: Advanced Concepts
If you’re an IT professional, you already know that new architectural patterns and concepts are changing the way we design for the cloud. This is especially true in today’s remote and hybrid work environment, where choosing the right type of cloud platform can determine the success and functionality of your entire IT enterprise.
Join instructor Lee Atchison as he shows you the business and tech-related basics of orchestrating and managing an effective cloud architecture. Learn about the various types of cloud platforms and architecture types, as well as how to determine which option will work best for you. Explore service and microservice architectures, data management tools, serverless computing, and cloud security as you build out your infrastructure. Lee shows you how to successfully incorporate automation and management practices into your architecture strategy so you can create a more efficient environment for software development.
https://www.linkedin.com/learning/cloud-architecture-advanced-concepts-14595141
Generative AI with Large Language Models
https://www.coursera.org/account/accomplishments/verify/D2YYYONEF8PU
Awards & Achievements (1)
🏆 EMC Excellence Award
Description
Patents (1)
Optimized backup of clusters with multiple proxy servers
description
dd