Technology & Product Leader

Contactless IR Temperature monitoring - battery powered device (WiFi, Bluetooth)

Software, hardware and device prototyping of a contactless temperature monitoring device. ( links ) . The device was developed in multiple variants - based on EspressIF ESP32 (WiFi+Bluetooth), ESP8226 (WiFi only), STM32 (Wifi+eSim Cellular) and was a battery powered device which could perform non-contact temperature monitoring using an IR-Melexis sensor connected through an I2C/SPI interface to the main board. The project involved developing the application software to read from the IR sensors through multiple actuators or scheduled reads and transmit this data to a AWS IoT Core cloud service through MQTT. 

The cloud software was also developed to write the data from these devices to S3 and Dynamodb and show this data in real-time over a BI dashboard.  

Oct-2019 to Jun-2020

[Active support]

Locixx GmbH Personal Location Tracking

The personal tracker watch is battery powered and connected to a cellular network through a SIM. The tracker sends location information to configured clients through TCP at scheduled intervals or whenever there is an SMS trigger. The project involved writing a parser for the TCP packet arriving from the watch and reading the GPS information - such as latitude, longitude, strength, heading, number of satellites latched, etc. . This data was then saved to a cloud database (MySQL) to be read by an application upon search requests.  

Jun-2020 to Sept-2020

[Active support]

OBD-II Vehicle Location Tracking

The project involved end-to-end solution development for a location tracking and driving behavior analysis. The OBD device could read the vehicles CAN bus information as well as provide accelerometer and position data. The project involved integrating these fleet of devices to a cloud backend and aggregating data to be made available to an end application.  

Jan-2022 to present

Industrial IoT - OPC-UA PLC - Factory operational metrics and AMR metrics

The project involved 2 parts. 

• The first part was integrating industrial operations data from an automatic bottling plant through a PLC connected to an OPC-UA server. The cloud-app would read data from the OPC-UA server and then aggregate the operational metrics - such as plant downtime, total bottles processed in an hour/day, rejects count, OEE, etc. 

• The second part was collating data from Omron Autonomous Moving Robots and feeding these to the operational dashboards to monitor the efficiency and usage of the AMRs. 

Jan-2022 to present

Embedded Software Projects:  Multimedia codecs optimizations and Media systems development on ARM Cortex and TI-DSP based processors

These were multiple long duration projects spread over multiple years. 

Multimedia Codec Optimizations: Codecs such as MP3, H264 are typically written in C/C++ language. To run these codec softwares on low-power devices such as those enabled by ARM-Cortex-A8 or TI-DSPs needs both compute and memory optimizations to enable real-time encoding or decoding. Understanding the processor architecture and Memory I/O are key to determining the optimization approach. 

System development: Development of Audio/Video playback systems, Video-VoIP communication systems on ARM and TI based processors. Have worked on optimizing components, memory/cache optimizations, system integration on ARM7/9/9E/11 and TI-C64x processors. Competent with ARM and TI-C64x+ platforms. OpenMAX IL, Gstreamer, TI Media framework, TI BIOS, TI Code composer studio, ARM-linux, ARM ADS, linux makefiles 

2003 to 2010

Projects

Description

Time 

General Management @ Vodafone Idea Ltd. 

This was a greenfield project for the largest Indian Telecom - Vodafone Idea Cellular Ltd. It involved setting up an OTT platform from scratch including managing RFP and RFQ processes, multi-vendor evaluations, handing out projects and services to various vendors and managing their deliveries and output, working with commercial and product teams to make progress.   

Vodafone and Idea underwent a merger and I was responsible for technology and live-service merger of Telecom Value Added Services and Digital Services. 

I handled Data Asset Monetization programs with multiple vendors in advertisement, finance-credit-rating, weather-forecasting and similar verticals who leveraged Machine-Learning over telco data for improving services. 

July 2016 to Sep 2019

Product Management @ Mediamelon Inc.

Started from a PoC and few patents to defining the product and implementing it in a way that it can fit the industry offering. Worked with multiple industry folks and CTOs of partner companies to create and publish joint collaterals and define the product in a way that it becomes easy to adopt. Deeply involved with hiring and bringing the right folks into the organization. 

June  2013 to Aug 2016

Engineering Management @ 

Ittiam Systems Pvt Ltd

I headed a team of around 10 engineers working on developing various audio and video systems, compression software products and mobile applications. Was responsible for hiring, on-boarding and  building human-resource capabilities and team-management. 

Jun-2008 to May-2013

Projects

Description

Time 

Analytics Data-Platform - Data-streaming-stack, API backend, React-based frontend -

ConnectedThinks Pvt Ltd

This is a Industrial data platform capable of connecting to multiple machinery (through OPC-UA, MQTT or REST APIs), transform data (through Kafka, Kafka streams and KSQLdb stack) and provide data-insights (through a home-grown BI tool capable of real-time monitoring) and connection to other BI tools such as Superset and Looker through presto/trino. The stack is built and deployed through docker, docker-compose and swarm and capable of on-prem or on-cloud deployments.  

2019-Dec to present

 [Active Development & Support]

OTT Video Service Analytics & Real-time anomaly detection

Mediamelon, Inc

The project involved architecting and implementing a real-time system which can work with real-time messages of the high-order and frequency and is able to detect anomalies in real-time. The size was of the order of 1k to 3k messages per second where each message has 100+ parameters. Kafka and KSQL were utilized to implement this system and custom KSQL UDF functions were written which could run ML prediction models to predict the anomalies. 

2021-June to 2021-Dec