Akhil Mahajan

AI Researcher

Bengaluru, Karnataka, India6 yrs experience

Most Likely To SwitchHighly Stable

Key Highlights

Reduced user delinquency rates significantly.
Developed large-scale machine learning models.
Mentored junior data scientists effectively.

Stackforce AI infers this person is a Data Scientist specializing in Fintech and Machine Learning.

Contact

akhil123ag@gmail.com LinkedIn

Skills

Core Skills

Machine LearningData Science

Other Skills

AlgorithmsAmazon AthenaAmazon S3CC++Computer ScienceDatabricksDeep LearningGraph Neural NetworkInformation RetrievalLSTMLaTeXLinuxNumPyPandas

About

Senior Data Scientist with 4+ years of experience in the Fintech and BNPL domain. Skilled in building and deploying large scale models from scratch that effectively reduced delinquency rates and helped in driving growth. Experienced in creating dashboards and alerts to monitor model performance and translating complex data into actionable insights.

Experience

Simpl

2 roles

Senior Data Scientist

Promoted

Apr 2023 – Present · 2 yrs 11 mos

Setup Simpl's first Graph Neural Network (GNN) model which helped reduce Simpl’s repeat user delinquency with a precision uplift of 2% and recall uplift of 1.5% and an annual savings of ₹15.6 million.
Setup Automated Databricks workflows for network supernode cleanup, user unblocking and sending comms; proactive communications helped improve overall growth and NPS score.
Setup Dashboards in Databricks and alerts using Zenduty to monitor GNN performance and get insights on daily users unblocked after sending various communications.
Mentored data scientists on various projects, helping with knowledge sharing and deployments.

Graph Neural NetworkDatabricksZendutyMachine LearningData Science

Data Scientist

Jan 2022 – Apr 2023 · 1 yr 3 mos

Built and Deployed Simpl's first deep learning model (LSTM) which helped reduce new user delinquency from 28% to 22% with an annual savings of ₹57 million.
Managed and preprocessed over 9 billion records of approval data to deploy the model.
Created a MaaS(Model as a Service) which currently handles over 60 million approval calls/day and blocks potential fraud users in real-time.
Optimized workflow execution by 85% through PySpark optimizations, reducing processing time from 7 days to 1 day.
Setup dashboard in Databricks and alerts using Zenduty to monitor the performance of the model.
Languages used - SQL, Python
Frameworks/Libraries used - Tensorflow, s3fs, Pandas, Numpy, Matplotlib
Tools used - Sagemaker, Cloudlift, CircleCI, Dask, Athena, S3, Pycharm

Deep LearningLSTMPySparkDatabricksMachine LearningData Science

Indraprastha institute of information technology, delhi

2 roles

Postgraduate Researcher

Jan 2021 – Aug 2021 · 7 mos

Worked on a Capstone Project in the field of Machine Learning under the guidance of Dr. Anubha Gupta at IIITD Signal processing and Bio-medical Imaging Lab. The aim of this project was to identify if a healthcare worker(HCW) involved in Covid-19 treatment was stressed based on his/her ECG data only. This work has been published in IEEE Xplore.

Teaching Assistant

Aug 2020 – Jan 2022 · 1 yr 5 mos

TA for Computer Networks, Systems Programming, Linear Optimization, Money and Banking and OOPD courses.

Emc

Student Ambassador

Dec 2013 – May 2014 · 5 mos · Chennai Area, India

To act as an interface between VIT university and the EMC organisation and organise various educational seminars in colleges.

Csir-indian institute of integrative medicine (csir-iiim)

Research Intern

May 2013 – Jul 2013 · 2 mos · India

Completed 4 weeks of summer research internship where the major focus was to learn about various techniques used in data cleaning, data analysis and applying predictive modelling on data collected from various medical stores.