Aug 2019 – Feb 2021 · 1 yr 6 mos · Greater Bengaluru Area · On-site

• AI-Driven Telematics and Driver Assistance System Development
• Led the design and implementation of an intelligent telematics platform for enhanced driving safety and user interaction. Key innovations included wake word detection ("Hello Kruzr") for hands-free voice activation, a Q&A bot for triaging incoming calls to minimize distractions during driving, and geofencing alerts for accident-prone zones to proactively notify users.
• Advanced Telematics System: Utilized Inertial Measurement Unit (IMU) sensors with sensor fusion filtering techniques (e.g., Kalman filters) to accurately detect and analyze driving events such as hard acceleration, braking, sharp turns, and lane changes, improving real-time risk assessment and response accuracy by integrating multi-sensor data streams.
• Deployment Impact: Deployed on edge devices, handling high-volume inferences to support safer driving experiences, aligning with my expertise in on-device AI optimizations.
• This system has contributed to reducing driver distractions and enhancing road safety through seamless AI integration.

May 2017 – Aug 2019 · 2 yrs 3 mos · Bengaluru, Karnataka, India

• Apparel Recommendation System: Designed an advanced apparel recommendation engine leveraging visual similarity algorithms and deep learning-based feature extraction. Enabled personalized user experiences by integrating text-to-image query processing using multimodal embeddings, resulting in a significant boost in user engagement and conversion rates.
• Article Recommendation System for Healthcare Portal: Developed a cutting-edge recommendation system for an online healthcare platform, employing topic modeling (LDA), headline sentiment analysis with NLP, and graph theory-based network analysis. Enhanced content relevance and user retention by delivering highly targeted articles, driving a notable increase in session duration.
• Media Campaign Analytics for Prescription Lift: Architected a robust media campaign analytics system to quantify prescription lift across 50+ drugs. Utilized attribution modeling and time-series analysis to measure campaign effectiveness, empowering stakeholders with actionable insights that optimized marketing spend and improved ROI.
• Default Prediction Model for Real Estate: Built a high-accuracy default prediction model for a real estate firm, integrating past transaction data, user demographic attributes, and complaint sentiment analysis. Applied ensemble learning techniques (XGBoost, Random Forest) to enhance predictive power, reducing financial risk by identifying at-risk accounts with precision.
• Sales Rep Campaign Design Tool: Engineered an intuitive campaign design tool for sales representatives, incorporating predictive analytics and optimization algorithms. Streamlined campaign planning processes, enabling data-driven decision-making that boosted sales team efficiency and campaign performance metrics.

May 2016 – Jul 2016 · 2 mos · Bengaluru Area, India

• Demand Forecasting for European MVNO: Engineered a sophisticated demand forecasting model utilizing Long Short-Term Memory (LSTM) networks within a recurrent neural network (RNN) architecture, implemented using Theano for efficient tensor operations and automatic differentiation. This deep learning approach captured temporal dependencies in sequential data, enabling precise prediction of network usage and inventory needs, which delivered annual savings of approximately 10 million Euros—equating to 7% of total operational costs—by optimizing resource allocation and reducing overstock.
• National-Level Retail Store Clustering: Developed an advanced clustering system for retail stores across a national footprint, applying non-linear dimensionality reduction techniques such as Conditional Restricted Boltzmann Machines (C-RBM) built with Theano to handle high-dimensional feature spaces and uncover latent patterns. Integrated deep generative modeling and contrastive divergence for training, resulting in optimized store segmentation that enhanced supply chain efficiency, targeted marketing strategies, and overall revenue growth through data-driven insights.

Dec 2014 – Dec 2014 · 0 mo · India