Alok Kumar
Cloud Engineer at Hewlett Packard Enterprise.
3.1
Years of Experience
Education
university of massachusetts amherst, maharaja agrasen institute of technology, ramkrishna mission vidyapith deoghar
Companies
hewlett packard, university of massachusetts amherst, hewlett packard, university of massachusetts amherst, deepmind, university of massachusetts amherst, goibibo, goibibo
Reach out to Alok Kumar via Email, InMail and SMS drip
by installing Chrome extension
Alok's contact details
Email (Verified)
aloXXXXXXXXXXXXXXXom
Mobile Number
+14XXXXXXX11
Github
aloXXXmv
Experience
2023 - Present
hewlett packard
Cloud Engineer
I am working as a Cloud Software Engineer at HPE in Storage Fabric team.
2022 - 2022
university of massachusetts amherst
Graduate Teaching Assistant
Worked as Grader for 490A (Applications of natural language processing)
2022 - 2022
hewlett packard
Cloud Software Engineer Intern
I worked as a Cloud Software Engineer Intern at Hewlett Packard Enterprise. During my internship, I worked on some interesting and challenging real-world projects. 1. I designed a system capable of monitoring the liveliness of on-premise docker containers from cloud, capable of detecting and restarting any unhealthy container. The system reduced container downtime significantly. Tech: Java, Golang, kubernetes, Docker, Kafka, REST API, FAST API, RabbitMQ Server 2. Integrated Prometheus client into various microservices across storage fabric manager(SFM) to collect four golden monitoring signals(Latency, Traffic, Error and Saturation). Created Grafana Dashboards to visualize the golden signals and setup alerts for any unexpected behavior. The system reduced downtime of microservices all across SFM by ~10%. Tech: Prometheus, Grafana, Python, Golang, kubernetes, Docker, Bastion
2022 - 2022
university of massachusetts amherst
Student learning assistant
Worked as Student learning assistant for umass Athletics department.
2022 - 2022
deepmind
Graduate Student Researcher
UMass CS696DS Industry Mentorship Program Research Project Name: Random Block Sparse Networks - benefits of width without its costs. 1. Explored if a random block sparse wide neural network can match the performance of the wider network but be more efficient,to verify the conjecture if width is primarily responsible for improved performance. 2. Performed numerous experiments on Fully connected neural networks, Convolutional Neural Networks and Residual Neural Networks for different width and sparsity ratio. 3. Found that the accuracy of model in all cases increased as we increase the width and sparsity by a similar factor, which is in support of our conjecture Industry Mentor: Manzil Zaheer (DeepMind, Google)
2021 - 2023
university of massachusetts amherst
Master of Science in Computer Science
2019 - 2021
goibibo
Software Engineer
I worked as Software Engineer at Goibibo train's backend team. I was mainly responsible for managing and expanding server-sided systems and client-facing APIs. I was primarily using Golang and Django as my tech stack. Apart from that I also used other additional tools and technology like Git, Gerrit, Docker, Kafka etc. At Goibibo I worked on multiple customer-facing projects which impacted more than 100,000 customers on daily basis. A couple of worth mentioning projects that I did here during my stint were: 1. Designed a graph database and an optimized search query algorithm to query data from a dataset of 7000 trains and 10,000 routes with an average complexity of O(n). 2. Developed a highly optimized recommendation engine capable of suggesting the most optimal route for the commuter in case of non-availability of original booking request based on variables like nearby stations, availability, price, etc in ~300ms over the combination of ~10,00000 routes. 3. Developed a high-performance e-ticketing platform for the Hyderabad subway system catering to 1 Million+ commuter per month. 4. Developed a central system to facilitate the exchange of data among various teams at Goibibo which was used for designing learning and prediction systems. The system comprises multiple central APIs which were integrated by teams all across Goibibo to send different types of data to the central data repository. The system also provided another set of APIs to the data team using which they were able to consume the data stored in the central repo for training ML models used for various kinds of forecarsting. The system was quite stable with an average response time of ~15ms at a scale of ~130k rpm.
2018 - 2018
goibibo
Back-End developer intern
I worked as a software developer intern at Goibibo Inventory management team. As an intern, I worked on the project of developing a dashboard-enabled solution that powered & monitored most of the A/B tests and experiments being served at Goibibo. This empowered product & business teams with better visibility which in turn led to data-driven planning & execution of critical initiatives. For this project, I created an intuitive dashboard in Reactjs that enabled the people in leadership positions to visualize the results of various A/B tests going across the Goibibo. I also created the backend system written in Golang on top of the Facebook Planout experimentation framework. This system was used both for powering the test cases and automatically generating analysis reports from the experiment results. For my above work as an intern, I was awarded a pre-placement offer from Goibibo.
Experience
96 Skills
Algorithms
Algorithms
Android Development
Application Programming Interfaces (API)
Backend
Bootstrap
c
C++
Computer Vision
Computer Vision
Computer Vision
Core Java
CSS
Data Science
Data Scientist
Data Structures
Deep Learning
Deep Learning
Deep Learning
Django
Django
Django
Django REST Framework
Django REST Framework
Django REST Framework
docker
docker
Docker
Gin
Gin
Git
GitHub
Go (Programming Language)
Go (Programming Language)
golang
golang
Golang
Grafana
HTML
Java
JavaScript
kafka
kafka
Kafka
Kubernetes
kubernetes, Docker, Kafka, REST API, FAST API
kubernetes, Docker, Kafka, REST API, FAST API
Linear Algebra
Linear Algebra
Linux
Microservices
MongoDB
MongoDB
MongoDB
MySQL
Natural Language Processing (NLP)
Neural Networks
NumPy
NumPy
NumPy
Object-Oriented Programming (OOP)
Operating Systems
pandas
PHP
Programming
Prometheus
Python
Python (Programming Language)
Python (Programming Language)
PyTorch
PyTorch
Pytorch
RabbitMQ
reactjs
reactjs
Redis
Redis
Redis
Representational State Transfer (REST)
Research Scientist
REST
REST API
Search
Software Development
Software Engineer
Software Engineering
Spring
SQL
SQL
SQL
storage
teaching
TensorFlow
Vision
Web Design
Web Development
Education
2021 - 2023
university of massachusetts amherst
Master's degree
Computer Science
2015 - 2019
maharaja agrasen institute of technology
Bachelor of Technology - BTech
Computer Science
2007 - 2012
ramkrishna mission vidyapith deoghar
CBSE 10th board
Physics
