Oct 2020 – Feb 2021 · 4 mos · Bengaluru, Karnataka, India

• Machine learning Architect for handling data projects of HTIPL Big Data team.
• Managing data pipelines and end to end data engine architecture for UniBusiness platform (For small businesses) for connecting merchants and buyers. Work mostly into creation of machine learning solutions for search and targeted ADs.

Apr 2018 – Oct 2020 · 2 yrs 6 mos · India

• Stealth Mode Technology Startup

Mar 2017 – Apr 2018 · 1 yr 1 mo

• Expertise in Project Management, Data Analytics, algorithms and data modelling involving Signal Processing, Feature Extraction, supervised and unsupervised Machine Learning and AI techniques.
• o Algorithm design using data analytics on sensor data. Sensor data signal processing, heuristics and machine learning techniques are used to build comprehensive solutions to problems like – Automatic Gym Activity recognition and repetition counting, Blood Pressure estimation on smartphones and wearables, mental stress detection, sleep quality assessment and apnea analysis etc.
• o Built an automatic activity recognition engine and Blood Pressure estimation engine deploying smart phones and wearable gear device data using machine learning techniques (ANNs and DNNs respectively).
• o Commercialized applications “StayFit+” for automatic activity recognition and repetition counting on Tizen platform and "Insta BP" on Android for non invasive smartphone based blood pressure monitoring. “StayFit+” has over 100K downloads on the Samsung Galaxy App store.
• o Worked on end-to-end prototyping and model development of "InstaBP", a blood pressure monitoring app for Samsung's S and Note series smartphones. Performed data processing, feature engineering and trained LASSO regression models and Deep neural network models on blood volume signals obtained from the heart rate sensor using Tensor Flow, NumPy, SciPy and scikit-learn in Python.
• o Outcomes: a) A1 US Patent on Blood Pressure estimation. b) 1 Publications in IEEE EMBC 2016 and 2018 with more than 20 citations. c) InstaBP was declared a winner in NASSCOM's AI Game Changer Awards. c) 30000+ downloads on Google Play Store.

Jul 2015 – Feb 2017 · 1 yr 7 mos

• Expertise in Data Analytics, algorithms and data modelling involving Signal Processing, Feature Extraction, supervised and unsupervised Machine Learning and AI techniques.
• o Building software oriented solutions to futuristic problems and products to come up in health and fitness domain.
• o Algorithm design using data analytics on sensor data. Sensor data signal processing, heuristics and machine learning techniques are used to build comprehensive solutions to problems like – Automatic Gym Activity recognition and repetition counting, Blood Pressure estimation on smartphones, mental stress detection, sleep quality assessment and apnea analysis etc.
• o Built an automatic activity recognition engine and Blood Pressure estimation engine deploying smart phones and wearable gear device data using machine learning techniques (Artificial Neural Networks and Deep Neural Networks respectively).
• o Commercialized applications “StayFit+” for automatic activity recognition and repetition counting on Tizen platform and "Insta BP" on Android for non invasive smartphone based blood pressure monitoring.
• o Worked on end-to-end prototyping and model development of "InstaBP", a blood pressure monitoring app for Samsung's S and Note series smartphones. Performed data processing, feature engineering and trained LASSO regression models and Deep neural network models on blood volume signals obtained from the heart rate sensor using NumPy, SciPy and scikit-learn in Python. Deployed the entire code on Java for developing the Android backend.
• Outcomes: a) A1 US Patent on Blood Pressure estimation. b) 1 Publications in IEEE EMBC 2016 and 2018. c) InstaBP was declared a winner in NASSCOM's AI Game Changer Awards. c) 30000+ downloads on Google Play Store.

May 2014 – Jul 2014 · 2 mos · Greater Bengaluru Area

• User Context Detection and Route Suggestion
• o Inferring User Modes : Walking, Running, Driving, Public Transport using supervised ML algorithms
• o Simple live Android Application for User Mode Detection – Walking, Running and Driving
• o Identifying Trips, aggregate level statistics and Estimated time of arrival using Open Street Maps
• o Determining home, work location and other places of interest
• Analysis using several supervised and unsupervised ML algorithms making use of Sci-Python, MongoDB etc.
• Received a Pre-placement offer for a full time employment at SRIB.