Dec 2022 – Jan 2024 · 1 yr 1 mo · Bengaluru, Karnataka, India

• Kromium is a Creator Marketing Protocol that aims to use smart contracts to revolutionize how stakeholders interact with each other in the creator economy.
• We aim to use web3 to construct a p2p adtech model that makes creator marketing more transparent, seamless, scalable and enforce standardizations that can finally value creators in a fair manner.

Apr 2022 – Nov 2022 · 7 mos · Bengaluru, Karnataka, India

• Experimented with several interesting problems at the intersection of computer-vision, digital marketing and Data Science. You can find some of our work at https://github.com/LaotechLabs .

Mar 2021 – May 2022 · 1 yr 2 mos

• Lead Instructor at Launchpad's Machine Learning Bootcamp at Levi Strauss & Co. for AMA (Asia, Middle East, Africa) region.
• Consultant/Mentor to multiple projects in QA, Audit, and Retail teams for the AMA region.
• Led a team of 20 students to build a Loyalty prediction model for Europe region.
• Worked towards building a model that can predict the revenue of a store using demographic and locational information.
• Led a team of 5 to build a SOA model to predict Cotton Commodity prices and understand trends using live news data.

Jul 2020 – May 2022 · 1 yr 10 mos

• Lead Instructor at Launchpad's Machine Learning Bootcamp at Levi Strauss & Co. for AMA (Asia, Middle East, Africa) region.
• Led the development of a custom similar & complementary product recommendation solution for Dockers (https://us.dockers.com/).
• Built a multimodal visual-textual search engine to browse through a catalog of
• 250,000 clothing items.

Jan 2020 – Jun 2020 · 5 mos

• Built a system for WP Diamonds that leverages recent advances in computer vision and NLP to effectively price second-hand watches and jewelry objects.

Sep 2019 – Dec 2019 · 3 mos

• Highly competitive 4-month fellowship for machine learning and deep learning with an acceptance rate of around 6 percent.
• 1) Built a deep learning application capable of identifying stained(defective) furniture objects and walls. The application used a combination of GradCam based segmentation, RoomNet and YOLOv3 object detection for detecting defective furniture objects and stained walls.
• 2) Contributed towards the improvement of an interior design computer vision application. This application was built around a ResNet34 classifier capable of identifying 40 different outdated(or ”faux pas”) design trends.
• 3) Worked towards reducing the latency of a Video Analytics API by using video summarization through keyframe extraction.

Jan 2019 – Feb 2019 · 1 mo · Bengaluru Area, India

• 1) Was among the 60 fellows selected to pursue the highly selective Policy in Action fellowship with YLAC.
• 2) Worked towards creating a startup policy for Delhi that encouraged clean-tech, social-tech and civic startups.
• 3) Submitted a policy paper titled ”Creating a Startup Policy for Delhi” to the Government of National Capital Territory of Delhi.

Jan 2018 – Aug 2019 · 1 yr 7 mos · Bengaluru Area, India

• 1) Implemented an Ingestion API that ingests SDK logs data from over a million mobile devices on a daily basis and stores it in a manner that enables large scale analysis. Capacity: 500 QPS.
• 2) Built a disaster recovery data pipeline that could backfill an entire day’s data(approx 500 GB) from cold storage(S3) into the Elasticsearch data warehouse within an hour.
• 3) Mentored an intern on a project that involved setting up a Google BigQuery data warehouse for business analytics purposes. This project involved creating data pipelines that pulled data from Elasticsearch, Salesforce and Atlassian JIRA.
• 4) Part of the 5 member team that led the efforts to integrate Zendrive’s technology into Verizon’s telematics applications, some of which have over 10 million users.
• 5) An integral part of a massive undertaking to restructure Zendrive’s whole technology in a manner that complies to various data privacy regulations across North America and Europe.
• Received an award from Zendrive’s senior leadership as recognition for the above efforts and was declared ’The Zen of Getting Shit Done’.
• Technologies used: Python(Flask), Pyspark(Apache Spark), Elastic Stack, Postgres, Apache Thrift, Apache Parquet.
• AWS: Elastic Beanstalk, EC2, DynamoDB, RDS, S3, Route53, IAM, VPC, Secrets manager, etc.

Jul 2017 – Dec 2017 · 5 mos · Bengaluru Area, India

• Built a log analytics framework that can be used to monitor Zendrive’s system architecture and derive insights into the nature of traffic that enters the system. This framework was designed to handle 1 million events per minute.
• Technologies involved:
• Servers: Apache + Flask(Python) + mod_wsgi, nginx
• Elastic Stack: Filebeat, Logstash, Elasticsearch, Kibana
• AWS: EC2, Elasticsearch service, Elastic Beanstalk, S3, Route53, Elastic Load Balance, Cloudwatch, RDS
• Linux: Systemv, supervisord
• Load Testing: Locust

Dec 2016 – Jan 2017 · 1 mo · Bengaluru Area, India

• Worked towards integrating the docker ecosystem to an IoT based framework of ARM devices(RaspberryPi, CHIP) providing music solutions to various Pubs and Resto Bars across Bangalore
• Explored various cloud orchestration technologies such as Docker swarm and Rancher to orchestrate the functioning of various ARM devices from a centralised location.
• Used Docker swarm to orchestrate clients by creating overlay networks
• Technologies involved:
• Docker ecosystem(Docker Engine, Docker Compose, Docker Swarm), Rancher, AWS EC2 services, Linux

May 2016 – Sep 2016 · 4 mos · Madgaon Area, India

• Led a team of 10 students to implement the concept of Startup Expo where startups are given an opportunity to showcase their products/services in front of students and VCs that visit the fest.

May 2016 – Jun 2016 · 1 mo · Bengaluru Area, India

• Built a fully functional Android App from scratch for the Community Policing team comprising of the Bangalore City Police and individuals from Janaagraha
• Aim of the app:
• º Managing the workflow of community policing activities
• º Parameterising the community policing activities and generating data for further analysis.
• º Storing this data and building a framework to retrieve it in a form readable by laymen(Excel)
• º Enabling real time monitoring of users by keeping a track of their GPS coordinates
• Technologies involved:
• Client end: Android- Java + SQLite + FirebaseUI(User Authentication)
• Backend: Firebase(storage+auth), PHP/HTML/CSS/Bootstrap(For GPS Tracker interface and its Authentication)
• Retrieval: Python scripts to retrieve information from firebase and present in excel format

Dec 2015 – Feb 2017 · 1 yr 2 mos

• Worked with the following NGOs/non-profits:
• 1) Bhumi(Chennai) : Negotiated the development of a cheaper CRM software than the one currently in use. Suceeded in finding alternatives at 30-60% of the target budget.
• 2)A Friend Abroad: Developed and implemented the market entry strategy of https://afriendabroad.com/ for its entry into India.