Lohith G N
VP - Data Science
Bangalore
Summary
Data Science Leader with 14+ years of experience in AI, ML/DL, Gen AI and Agentic AI across Telecommunication, Banking, Retail, and Healthcare. Currently with the Data and Innovation team at NatWest Group, leading data science projects from concept to deployment.
Experienced in managing teams, driving product enhancements, planning releases, and engaging with customers to align solutions with business goals. Skilled in scalable AI deployment, LLMOps, and Gen AI consulting across global projects.
Hold an M.Tech in Data Science and Engineering from BITS Pilani and a PGPDM in Advanced Machine Learning from the University of Chicago. Passionate about using AI to drive innovation and business transformation.
Overview
14
14
years of professional experience
Work History
VP - Data Science & Innovation Team
Natwest Group
02.2025 - Current
- Led cross-functional teams (ML, DE, SE, Gen AI gateway platform) to deliver three major wealth and retail projects, achieving 80% higher efficiency and client satisfaction.
- Directed four successful release cycles, including one for product enhancement, two for bug fixes, one for GPT model upgrade and a hotfix for a automating data fetch for the project operation monitoring.
- I presented a poster and live demonstration on Autonomous Multiagent for detecting RCA for poor-performing performance metrics at the DSEC Forum in London, receiving strong, positive feedback for the concept.
- Collaborated with senior management to define long-term strategies and streamline AIOps processes, including handover checks, Agile progress tracking, and Gen AI model upgrade monitoring. Additionally, I led SteerCo calls to align team efforts and drive operational excellence.
Principal Data Scientist
Randstad Digital
01.2024 - 01.2025
- At Airbus France, I led the development of an intelligent chatbot that handles diverse aircraft domain data, utilizing technologies like Graph DB, Metaflow, and LLAMA3. This scalable solution improved component analysis and knowledge retrieval, boosting efficiency by 40% and accuracy by 30%, supporting over 100 use cases.
- For Synerion, I designed a test case generation tool using Claude 3.5 and Amazon Bedrock, applying LLMOps to cut generation time by 50% and improve reliability by 25%, accelerating client delivery.
- Developed a long-form video generation system for Airbus avionics installation guides, employing EasyAnimate, diffusion generators, and VAEs to enhance documentation.
- At CVS Healthcare, we created a Visual RAG system for ultrasonic image abnormality detection using Qwen2 and Colpali, achieving 90% diagnostic accuracy.
Role -Data Scientist -Gen AI, ML & Cloud
Rakuten Symphony
1 2022 - 01.2024
- Real-Time network analysis (Gen Ai in Rakuten Cloud): Designed and built low-latency LLM companion bot with over 82% accuracy to analyze highly utilized network cells with high PRB utilization. Automated daily insights across 10,000+ towers in mobile network, delivering performance updates and aligning strategic roadmap planning with senior management's vision.
- Document Data Retrieval with GenAI (Multivendor Integration): Engineered a custom LLM chatbot for extracting insights from multivendor documents. Integrated advanced NLP techniques like extractive QA, hybrid retrievers (BM25 & REALM), and Farm Reader (ROBERTa) with RAG methodology to enhance generative accuracy and reduce hallucinations. Successfully deployed the system for 30+ domain experts, improving user comprehension and decision-making by 50%.
- Network Performance Optimization Recommendation System: Led the development of a recommendation engine for underperforming mobile network cells, employing collaborative filtering and content-based modeling. Achieved an 80% reduction in proactive network measurement times, streamlining optimization efforts and improving operational efficiency.
- LLM Quality Enhancement Framework: Enhanced LLM reliability and safety through advanced monitoring and mitigation techniques, including selfCheckGPT, Toxigen models, and NER algorithms. Reduced issues such as hallucinations, data leaks, and toxic output by over 90%, ensuring regulatory compliance and robust system performance.
- Analytical Dashboard Development: Delivered 11 cutting-edge analytical dashboards leveraging techniques like Pearson correlation analysis, Apriori algorithms, and statistical distributions. Automated workflows using Apache Airflow, achieving a 90% reduction in network analysis time and accelerating data-driven insights for strategic initiatives.
Role - Data Scientist -NLP & Text Analytics
ITC Infotech
07.2020 - 12.2021
- Document Predictive Model for Claims Automation (UHG): Developed an industry-specific predictive model leveraging DistilBERT from Hugging Face to automate the classification of appeals and recalls in claim documents. Integrated Apache Kafka for seamless data ingestion, reducing manual effort by 70% and significantly optimizing claim processing for the UHG client.
- Probabilistic Model for Delivery Prediction: Designed and trained a Bayesian probabilistic model to predict the likelihood of C-section versus normal delivery. The model utilized blood reports, scanning data, and Social Determinants of Health (SDOH) as features, achieving an initial accuracy of 65%, paving the way for data-driven decision-making in maternal care.
- Cost Optimization with OpenAI CLIP Model: Improved medical report identification processes by implementing the OpenAI CLIP model, achieving a 30% reduction in operational costs. The solution was deployed on Azure Cloud, ensuring scalability and efficiency in medical workflow management.
- Chest X-ray Effusion Detection (POC): Conducted a comprehensive Proof of Concept (POC) for effusion detection in chest X-ray images. Trained a ResNet-50 model and employed advanced techniques such as ablation studies, Keras callbacks, and weighted cross-entropy strategies to enhance model performance. Reduced False Negatives and improved overall accuracy by 20%, delivering a robust solution for medical diagnostics.
Role - Data Scientist -ML & Cloud
Reliance Jio Infocom Limited
10.2016 - 07.2018
- Automated TAC Splitting with DBSCAN: Designed and implemented a DBSCAN clustering algorithm to autonomously split Tracking Area Codes (TAC) for overloaded regions, eliminating manual intervention. Leveraged MLOps techniques, including MLflow, for model tracking, deployment, and performance monitoring, ensuring consistent optimization of network efficiency.
- Hotspot Forecasting Model with SARIMAX: Developed a SARIMAX-based forecasting model to predict and identify 100 potential mobile network capacity hotspots nationwide. Incorporated MLflow for automated versioning, monitoring, and deployment, enabling seamless integration into network capacity planning workflows.
- 4G Network Capacity Monitoring Dashboard: Built a daily monitoring dashboard to track 4G network capacity, detect outlier cells using the Isolation Forest algorithm, and streamline operations through a web app. Implemented MLOps techniques with MLflow to monitor model performance, automate updates, and ensure reliable deployment of the solution.
Role - Senior Engineer -ML
ZTE Telecom India Pvt Ltd
01.2015 - 11.2015
- Predictive Modeling for Base Station Performance: Developed and deployed an exponential smoothing time series algorithm to forecast underperforming mobile network base stations. This proactive approach empowered the technical team to address issues in advance, improving network reliability and user experience.
- Automation Framework for Efficiency: Achieved a 50% reduction in manual effort by designing and implementing a Python-based automation framework, streamlining repetitive tasks and improving operational efficiency.
- Traffic Forecasting with Multivariate Time Series Analysis: Conducted advanced multivariate time series analysis using ARIMA and VAR models to predict daily, weekly, and monthly traffic patterns. Visualized peak traffic periods through intuitive dashboards, enabling data-driven capacity planning and resource optimization.
Role - Engineer -ML & Database
Ericsson
11.2012 - 12.2014
- Algeria - Marketing Analysis for Ooredoo: Applied Bandit Thompson Sampling and beta distribution to identify seasonal product categories and optimize pricing, enhancing targeted marketing strategies for 2 projects.
- Thailand - Predictive Maintenance for DTAC: Used K-means clustering and Additive Holt-Winters forecasting to predict downtime and maintenance needs, improving network reliability and efficiency.
Role - Senior Data Analyst
Nokia Siemens Network
04.2010 - 06.2012
Role - Mobile Network Business Analyst
ZTE Telecom India Pvt Ltd
07.2009 - 04.2010
Education
Master of Technology - Data Science And Engineering
Birla Institute Of Technology
Pilani, India 04.2001 -
PGPDM - Advanced Machine Learning
University of Chicago
Chicago 04.2001 -
BE - Electronics And Communications Engineering
SBMSIT
Bangalore 04.2001 -
Timeline
VP - Data Science & Innovation Team
Natwest Group
02.2025 - Current
Principal Data Scientist
Randstad Digital
01.2024 - 01.2025
Role - Data Scientist -NLP & Text Analytics
ITC Infotech
07.2020 - 12.2021
Role - Data Scientist -ML & Cloud
Reliance Jio Infocom Limited
10.2016 - 07.2018
Role - Senior Engineer -ML
ZTE Telecom India Pvt Ltd
01.2015 - 11.2015
Role - Engineer -ML & Database
Ericsson
11.2012 - 12.2014
Role - Senior Data Analyst
Nokia Siemens Network
04.2010 - 06.2012
Role - Mobile Network Business Analyst
ZTE Telecom India Pvt Ltd
07.2009 - 04.2010
Master of Technology - Data Science And Engineering
Birla Institute Of Technology
04.2001 -
PGPDM - Advanced Machine Learning
University of Chicago
04.2001 -
BE - Electronics And Communications Engineering
SBMSIT
04.2001 -
Role -Data Scientist -Gen AI, ML & Cloud
Rakuten Symphony
1 2022 - 01.2024
GitHub Profile
https://github.com/lohith0501
Skills
AI Technology: Agentic AI, Advanced Machine Learning (ML) Techniques, Data Mining & Predictive Modeling, Gen AI & Transformer Models, Statistical Analysis & Optimization, Experiment Design & A/B Testing, Time Series Analysis & Forecasting, Feature Engineering & Data Transformation, End-to-End Model Lifecycle Management, Big Data Analytics (Hadoop, Spark, Hive), Deep Learning Frameworks (TensorFlow, PyTorch), Data Visualization & Storytelling (Tableau, Power BI, Matplotlib, Seaborn), Large-Scale Data Pipelines & ETL Development, MLOps,LLMOPS & Model Deployment, Distributed Systems & Parallel Computing
Programming & Tools: Python, R,Langgraph, Langchain, SQL, NumPy, SciPy, Pandas, scikit-learn, MLlib, XGBoost, CatBoost, Jenkins, Docker, Kubernetes, Git, Bitbucket, NextJS
Cloud Platforms: AWS , Azure
Data Engineering: Apache Spark, Kafka, PostgreSQL, MySQL, NoSQL (MongoDB, Cassandra), Data Lakes, Hive, Hadoop, ETL Pipelines
Team Management: Production and Release planning,Team strategy planning, Resource budgeting, Hypothesis-Driven Analysis, Data-Driven Decision Making, Cross-Functional Team Leadership, Stakeholder Collaboration & Influence, Effective Communication & Storyboarding, Mentorship & Talent Development, Adaptability to Ambiguity, Creativity in Algorithm Development.
WHAT I AM PROUD OF
Poster Submission at DSEC London
- Multiagent Agentic RCA
Industry Project (The Math Company, PGPDM)
- Conducted advanced regression on US state price elasticity
Zomato Restaurant Ratings Prediction
- Ranked 5th in DLabs Data Science competition
International Onsite Support
- Provided in-person customer support in Shanghai, Shenzhen, Hanoi, Helsinki, Dhaka, Tel Aviv, Tunis, Bangkok, and Ho Chi Minh City
Certifications
Computer Vision Nanodegree Program by Udacity
Machine Learning Master's Program by Teclov
Extensive Python for AI by The School Of AI
Neo4j with LLM (Graph Database) by Neo4j