Abhishek Das

Senior Software Engineer

Delhi, India5 yrs 9 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

Expert in developing scalable systems for e-commerce.
Led AutoML framework development at Amazon.
Innovative solutions for complex data challenges.

Stackforce AI infers this person is a Backend-heavy Fullstack Engineer in the E-commerce sector.

Contact

Skills

Core Skills

JavaMicroservicesMachine LearningAws

Other Skills

Go (Programming Language)RedisAerospikeMongoDBPythonXGBoostDockerApache KafkaSpring BootAutoMLAmazon Web Services (AWS)MLOpsC++Data StructuresAlgorithms

About

Seasoned software engineer with over four years of experience in the dynamic e-commerce sector, having made significant contributions at industry leaders Amazon and Myntra. At Amazon's International Machine Learning team, I helped in the development of Automatic Machine Learning frameworks and ML Automation pipelines, collaborating closely with scientists to optimize model lifecycle processes. This role honed my expertise in Python, AWS services, Java, Spark, and XGBoost, equipping me with a robust toolkit for tackling complex data challenges. Currently, I'm leveraging my skills at Myntra as a key member of the Search team. Here, I'm dedicated to refining product recommendations, focusing on the intricate challenge of operating at Myntra's impressive scale. This role has deepened my proficiency in designing scalable systems, microservices, and working with databases and caches like Redis, Aerospike, and MongoDB. I thrive on the challenge of creating innovative solutions that can handle the immense scale and complexity of modern platforms.

Experience

5 yrs 9 mos

Total Experience

2 yrs 11 mos

Average Tenure

4 yrs 1 mo

Current Experience

Myntra

Senior Software Engineer

Apr 2022 – Present · 4 yrs 1 mo

Search: Currently part of Myntra Search team. The search team caters to understanding the query and then showing results specific to the user.
Major Features developed include:
1. User View Capping: Deboosting certain products which are not interacted by the user so that new products can surface and be viewed by the user
2. Trends Widget: Development of full width widget showing latest trends associated with the query the user has written. Example for shirts, trends might include "Boho Shirts" and products related to these are shown as part of widget
Communications Team:
The communication team sends notification to all the active users of Myntra userbase, Got understanding of how notification are sent to users at scale.
Furthermore took complete ownership of components responsible for showing Myntra Ads on 3P platforms like Google and Facebook.
Major Features developed include:
1. Reporting Dashboard: For showing reports of the notification campaigns executed by business team. The report covered the number of users, sends, failure along with the errors etc. This report could also be downloaded for further analysis
2. Creating Event Driven campaigns: This include the likes of Wishlist Low stock, Price drop etc and sending relevant notificaiton to the user. This lead to an increase in CTR of 8.4%
3. Creating New one time notification: This includes timer new versions of timer notifications, adding support for whatsapp notifications

JavaGo (Programming Language)RedisAerospikeMongoDBMicroservices

Amazon

2 roles

Software Development Engineer

Aug 2020 – May 2022 · 1 yr 9 mos · Bengaluru, Karnataka, India

Collaborating with AWS Science Team:
Onboarded Vowpal Wabbit algorithm as part of AutoGluon Library. Github PR Link: https://github.com/autogluon/autogluon/pull/1422
AutoML Framework
Took complete ownership of AutoML workflow being widely by internal amazon teams. This included business ops teams and ML Scientists
Features developed include:
1. Adding XGBoost as part of the AutoML framework
2. Adding support for HyperParameter optimization on a business metric
3. Adding support for Uncertainty and Diversity based Active learning strategies
4. Running 100s of parallel AutoML workflows from jupyter notebook and packaging the same as a library in less than 30 lines of code
AWS Monitoring Tool:
Tool was deployed using AWS Cloudformation which would monitor idle EMR instances and send notification to the creator of the cluster if it was idle for a given threshold.

PythonJavaXGBoostAWSMachine Learning

Software Development Intern

Jun 2019 – Jul 2019 · 1 mo · Bangalore, India

Pretraining and finetuning of BERT model from scratch and achieved a score of 97% accuracy on test data set
Packaged this model as a python library which was used by Amazon Scientists

PythonMachine Learning