Oct 2023 – Dec 2025 · 2 yrs 2 mos · Bengaluru, Karnataka, India · Hybrid

• Dual Usage Goods Classification:
• Led end-to-end development of a Dual Usage Goods classification system using Retrieval-Augmented Generation (RAG) and LLMs with Python, PostgreSQL (pgvector), and Langchain.
• Designed and deployed scalable data pipelines to extract export control regulations from PDF documents across 25+ countries and enriched the text using complementary internal data sources, significantly enhancing classification accuracy.
• Delivered a 25% improvement in solution accuracy and a 30% reduction in SME manual effort by replacing a legacy system with an advanced AI-driven solution.
• GIA Research Assistant:
• Designed and deployed a domain-specific RAG-based chatbot for Macroeconomics and Geopolitical Risk, automating 80%+ of SME queries reducing SME efforts and time using LLMs, LangChain, and PostgreSQL (pgvector).
• Built scalable multi-modal data pipelines (text/audio) and enabled geo-coordinate-based-queries, applied LLMs for speech-to-text and contextual chat handling, enhancing real-time system intelligence and user experience.
• Vessel ETA and Route Prediction:
• Developed and deployed vessel ETA prediction models using XGBoost and route prediction using CNN model, achieving a 15% performance improvement over the existing rule-based system.
• Designed a scalable pipeline in Python and PySpark to process large-scale maritime data and identified optimal shipping paths using graph-based algorithms.

Sep 2020 – Oct 2023 · 3 yrs 1 mo · Bengaluru, Karnataka, India · Hybrid

• Work Experience in Industries:
• 1. Chemical and Petroleum
• 2. Transportation and Logistics
• 3. Finance
• Some projects I've worked on:
• Petroleum Coke Quality Prediction/Optimization
• Predict Petroleum Coke Quality parameters and forecast 10+ crudes and 13 different properties time series using models such as XGBoost, Facebook Prophet, and Gradient Boost
• Optimized refinery process to achieve higher and sustained coke quality using Model Explainability (SHAP)
• Developed a robust Model Monitoring module that detected and prevented performance drifts; reduced model downtime by around 80% and possible a loss of around $0.6-1M/year
• Created end-to-end machine learning pipelines on Azure ML
• Cargo Vessel Trajectory and ETA Prediction
• Designed a network of Neural Network models using TensorFlow to predict the Vessel Trajectory and ETA. The network consists of 1000+ nodes and each node in the network consists of 2-3 different Neural Network models
• Created new features from online sourced data which improved model performances by 4-5% on average
• Conceptualized Solution performance criteria and implemented Model monitoring functionality
• Property Finance Content Intelligence
• Developed Bi-LSTM based Neural network using PyTorch to perform content intelligence on 14+ Tabular Data Formats extracted from PDF
• Incorporated LSH model and fuzzy matching to find the closest matching entity names; reduced search time by around 10 times
• The solution reduced the time to execute property finance evaluations by over 80%
• HSN Code Classification
• Developed BERT-based hierarchical classification Neural Network model using transfer learning; improved performance by 10%
• Extracted entities and relationships from product descriptions which are analyzed for auditing

Jan 2020 – Jun 2020 · 5 mos · Greater Bengaluru Area

• During my internship at IBM, I worked on the following assignment:
• Graph Analytics Platform:
• Implementing graph algorithms.
• Enhancing and improving the implementations for computationally efficient system.
• Integrating with Spark for cluster-computing.
• Created a modern, responsive and user-friendly web-platform.
• Technologies: JanusGraph, Gremlin, Spark, NodeJs, Python, jQuery, JavaScript.

Jan 2019 – Mar 2019 · 2 mos · Rajasthan, India

• During my internship at NIIT Limited, I worked on the following assignment
• Interaction Portal for Prospective students of the University:
• Login and Signup using OAuth.
• Developing features for effective interaction such as: Personal chat, Posts, Discussion Forums, User Profiles etc.
• Creating a modern, responsive and user-friendly web-platform.
• Technologies: PHP, Laravel, jQuery, JavaScript, HTML, CSS.

Dec 2017 – Apr 2018 · 4 mos · IIIT Delhi, New Delhi

• During my internship at NIIT Limited, I worked on the following assignment:
• SellBuyBook: An online platform to sell and buy books
• Created a modern, responsive and user-friendly web -platform.
• Login and Signup using OAuth process.
• Integrated payment gateways such as Paytm, RazorPay etc.
• Backend Development.
• Technologies: PHP, Laravel, jQuery, JavaScript, AWS Services (EC2, S3, Cloud9 IDE), HTML, CSS.
• KissanX: Crops and Agricultural tools Management Platform
• Created a modern, responsive and user-friendly web -platform.
• Backend Development.
• Technologies: PHP, Laravel, jQuery, JavaScript, AWS Services (EC2, S3, Cloud9 IDE), HTML, CSS.
• MetGiMet: Doctor Appointment Platform
• Created a modern, responsive and user-friendly web -platform.
• Login and Signup using OAuth process.
• Integrated payment gateways such as Paytm, RazorPay etc.
• Backend Development.
• Technologies: PHP, Laravel, jQuery, JavaScript, AWS Services (EC2, S3, Cloud9 IDE), HTML, CSS.

Jun 2016 – Nov 2016 · 5 mos · Rajasthan, India

• During my internship at NIIT Limited, I worked on the following assignment:
• Data Visualizer Platform
• Understanding and implementing different visualization techniques.
• Choosing suitable visualization for effective data representation.
• Creating a modern, user-friendly web-platform.
• Technologies: d3.js, jQuery, JavaScript, HTML, CSS.