Jan 2023 – Jun 2025 · 2 yrs 5 mos · San Francisco Bay Area

• Led the development of cutting-edge AI-powered revenue management systems tailored for the travel industry, leveraging advanced methodologies such as deep time-series forecasting and reinforcement learning to address complex pricing challenges.
• Designed and implemented off-policy reinforcement learning and contextual bandit algorithms to accurately estimate willingness-to-pay (WTP) and determine optimal bid prices, transforming airline revenue management strategies, and uplifting revenue.
• Conceptualized, designed, and developed a virtual feature store incorporating distributed engineering principles, responsible for providing model features during training and inference, reducing training times by 70%.
• Developed state-of-the-art multi-task and meta-learning approaches to train multi-objective deep learning models for forecasting and dynamic pricing. These advancements have empowered airline stakeholders with actionable insights, driving data-driven decisions and strategic planning.
• Led research initiatives through cross-functional collaboration. Contributions include database schema design, algorithm development, scalable ML system architecture, and seamless orchestration of training and inference pipelines.

May 2022 – Aug 2022 · 3 mos · Greater Seattle Area

• Interned with the Supply Chain Optimization Technologies (SCOT) organization during the summer, focusing on Network Flow Optimization.
• Developed an innovative algorithm combining machine learning and non-linear programming to strategically assign transship units within a supply chain network operating under limited capacity. The implementation of this algorithm led to a projected annual savings of $5M within the North American network.

Mar 2021 – May 2022 · 1 yr 2 mos · Dallas, Texas, United States

• Worked with Professor Lavanya Marla and Deepair Solutions on ML applications in airline ancillary pricing.
• ◦ Pioneered the development of an algorithm utilizing Subset Scanning and Causal Inference methods to detect data distribution shifts caused by unforeseen events like COVID-19. Achieved a notable accuracy gain of 16% compared to traditional methods for airline ancillary pricing.
• ◦ Engineered a groundbreaking method for dynamically pricing airline ancillaries in the face of distribution shifts. Incorporated a customized loss function, leveraging Causal Information and Deep Learning to model customer context, resulting in a 2% revenue uplift.

Oct 2019 – Jan 2021 · 1 yr 3 mos

• ◦ Successfully built and implemented a semi-automated Unsupervised Anomaly Detection system. Employed a diverse array of techniques including Long Short-Term Memory Networks (LSTMs), Convolutional Neural Networks (CNNs), Convolutional LSTMs, Expectation-Maximization (EM) algorithms, Mixture Models, and Time-Series State Classification. This system was adept at identifying real-time anomalies in manufacturing operations using IoT sensor data, achieving an F1 score of 0.8+ for two distinct clients in the manufacturing industry.
• ◦ Developed scalable and distributed pipelines in Kubeflow to train deep learning models across multi-GPU and cloud environments, improving training efficiency by 50% and reducing convergence time by 25%.

Oct 2018 – Oct 2019 · 1 yr

• ◦ Developed and deployed a state-of-the-art Deep-Survival Model to accurately predict the time-to-failure of manufacturing assets in real-time. This was accomplished by estimating the Weibull distribution and Cumulative Hazard function, with implementation enhanced by CUDA for improved computational efficiency. This approach significantly contributed to proactive maintenance planning, operational efficiency in manufacturing environments, and preventing impending machine failures on 2 different occasions.

May 2018 – Oct 2018 · 5 mos · Bengaluru Area, India

• Led the development of the Analytics Engine for Jupiter (an airline pricing intelligence system) including Database and schema design, data wrangling and model development including algorithms ranging from Machine Learning, Optimization and Game Theory.
• Managed a team of 3 engineers and built optimized data pipelines, scalable REST APIs, and machine learning models for real-time pricing insights, improving latency by 30% and ensuring >99% uptime.

Jun 2017 – Apr 2018 · 10 mos · Bengaluru Area, India

• Worked on building Jupiter, a Pricing Decision Intelligence System for clients across airline domain.
• Developed Triggers, a Predictive Engine and Automatic Alert System based on market fluctuation and irregularities, performance variation, events, competitor action and seasonality.
• Created a Pricing Recommendation System using Econometric Methods and Linear Programming and identified Nash Equilibrium to determine the dominant strategy in the market. The optimum pricing methodology helped in achieving a growth in revenue by 1%.
• Developed ETL connectors and Python scripts to pre-process, transform, and load data into MongoDB

Jan 2017 – Mar 2017 · 2 mos · Bengaluru Area, India

• Developed a model for the Simulation of market parameters using Adaptive Multivariate Regression and Splines to model the relationship between different parameters achieving an accuracy of 90%.
• Created methods to target individual customers with appropriate products by performing Customer Segmentation which helped in creating the right demand and increasing the market share by 10%.

May 2016 – Jul 2016 · 2 mos · Pune, India

• Worked in the Supply Chain Management of the Henkel Adhesive products produced in collaboration with the Asian Paints.
• Responsible for the Materials Requirement Planning of 8 SKUs having 49 different IDHs.
• Developed a structured framework for procurement and manufacturing resulting in the reduction of bottlenecks and increase in the overall sales by 20%.
• Coordinated with the various stakeholders for timely arrival of material at every end.

Jan 2016 – Apr 2016 · 3 mos · Roorkee, India

• Sangram is a 3-day long official sports festival of IIT Roorkee held every year in the month of April.
• Responsible for management of all activities held during the festival which included tournaments, ceremonies etc.
• Led a team of 30 student coordinators from 15 different sports.
• Volunteered for the planning of logistics of 2000+ student participants from various institutions of the country.

May 2015 – Jul 2015 · 2 mos · Bengaluru, India

• Worked in the area of Design for Sustainability under Centre of Excellence for Design of Sustainable Products, Services and Manufacturing Systems sponsored by Indo-US Science and Technology Forum (IUSSTF)
• Involved in the development of a tool which guides the designers in designing sustainable systems.
• Conducted the pilot study for validating the effectiveness of the tool.

Mar 2014 – Jan 2016 · 1 yr 10 mos · Roorkee, India

• Project involved design and fabrication of a Formula Style Electric Race Car which competed in FSAE Australasia held during Dec,2015 at Melbourne, Australia
• As a Mechanical Team member, worked on FEA methods for component design, fabrication methodologies and physical testing setups.