Experiences

AI systems Architect

ai4Good

Feb 2021 - Present,

AI for Good is driving forward technological solutions that measure and advance the UN’s Sustainable Development Goals. We create impact by bringing together a broad network of interdisciplinary researchers, nonprofits, governments, and corporate actors to identify, prototype and scale solutions that engender positive change. Founded in 2015 by a team of Machine Learning and Social Science Researchers in the US and Europe, AI for Good is headquartered in Berkeley, California with an international network of core team members, partners and volunteers supporting our work.

Responsibilities:

I Work on the aiforgoodsimulator which is a web tool for NGOs and local authorities to model COVID-19 and other epidemics inside refugee camps and prepare timely and proportionate response measures needed to flatten the curve and reduce the number of fatalities. This tool is designed to help humanitarian workers prevent the worst outbreak scenarios and design an optimal intervention strategy.
My responsibility is to build system with the best process and system optimisation techniques that are most required for the application.

Co-Founder

yknot-Concepts

March 2020 - Present,

Co-founded the team and currently heading multpile projects that focusses on human-computer interaction. Building open source projects and also designing architectures that helps people percieve artificial intelligence on competitive edge devices. Hence making it available for everybody on the face of this planet
Currently we are a small team of 10 Engineers , 1 graphic designer and a product designer.

Responsibilities:

thedesign.app (Under construction and is expected to release soon with white paper)
- Targetted for merchants and indepedent freelancers who own their own business and are trying to make progress online in this pandemic season. It facilitates a platform for the clients to place an order in their merchant site and also follow the progress of their order. It also helps to manage multiple freelancers with one supervisor to assign them tasks and monitor them.
- Although this first version of the project is aimed for free lancing graphic designers we will be making it opensource and will be releasing a framework where any independent professionalists like attornies, content-writers can modify content and use it like zammad.
- To make the adoption easier we will be releasing our white paper in arxiv and attempt to setup a forum with support team.
theGreetDoll
- This is a small AI-customised gift made of cardboard. When you a particular person is at the table it recognises them and wishes them.
- The best part about the greetDoll is that the manufacturing cost is less than 10$ and is independent of any cloud servers
- Configuration and Inferencing is done on the edge device.
- This also contains a easily configurable Digital-slam Book.
theRealSmartSwitch
- We are making a smart switch that does not require to rewire the entire building to make it smart. We are making a detachable hook-bot that runs a servers on its own and communicates with the inmates of the house through mobile app and voice assitants like Goolge and Alexa.
- The estimated manufacturing cost is less than 4$

AI-Systems Engineer, Instore AI and Customer Intelligence Team

Capillary Technologies

Dec 2019 - Present, Bangalore

Our products involve developing systems that involves Artificial intelligent solutions that improves Customer-Retailer relationship. This involves studying the customer counts, their demographics and their activity in the store and coming up with suggestions to improve shopping experience.

Responsibilities:

Five patents to be filed titled as - A contemporary approach for Industrial IoT software packaging, encryption and distribution
- An architecture for mass Industrial IoT devices management with more security and operation ease
- Synchronised co-processor architecture for retailer’s demography data acquisition
- IoT cum Vision-based solution for device localization in remote fields
- AI-Devices Assembly for person detection with superlative performance
Smart Store Team
- Design customer facing system architectures for edge computing devices to acquire data and communicate with the data hub.
- Involves system security and the way it self configures itself to send heartbeat, report anomality and smooth auto-installation of new packages if available in the deployer.
- Make ease of operations features.
- To build IoT based debugging features on the device. So that if the system goes down remote-debugging can be done. - Desiging a univeral inference engine that interprets any heavy Deep learning models on low compute edge devices.
Customer Intelligence Team
- Design and maintain AWS based system architectures to train and infer customer data using big data tools like hadoop, spark.
- Data Engineering on heavy datasets to make sure that analysis is happening smoothly without any stress on the inference engines.
- Also improve e-commerce based propensity models that recommends products or recommends the right promotion content or the promotion time slot for the target user to receive the message so he responds appropriately.

Strategic Investment Intern

Euro Exim Bank

Apr 2019 - Dec 2019, Reporting Office London

Delivered research and realtime projects on cutting-edge technologies. Also involves market research which to analyse the audacity of particular technologies for the bank to invest and build products.

Responsibilities:

Brain Computing Interface - Develop systems to tap EEG signals and export them to proper waveforms with relevant pre-processing. The acquired data was used for creating a feedback system for brain therapeutic machines that help us calm down and meditate.
Bone Conduction
- Develop minute spying devices that works on bluetooth and bone conduction transducers
- Develop bone conduction microphones that outperforms throat microphones at extremely high noise environments
Pizza Vending Robot - To design an industrial standard, robot ecosystem that receives customer orders, cook the pizza and deliver it to them. The system also included a rail bot that can brings the vendor bot the required elements on demand.

Data Analyst Intern

Einnel Technologies

Jun 2018 - Jul 2018, Chennai

Responsibilities:

Developed an interface for my team’s mechanical engineers that does multivariable-optimisation techniques on huge car-crash datasets

IoT Developer, Test-Developments team

Nokia Solutions and Networks

Apr 2018 - Nov 2018, Chennai

Responsibilities:

Designed and implemented ‘Universal Overall Engine effectiveness Module’ (UOEE module).
- Replaced NI-6501s that was very hard for the technicians to interact with.
- UOEE module cut several costs with respect to components and maintenance - The new system had UI that made the technicians life better 10x and gave a better perspective of the system.