Shubhasmita Panda
Bengaluru
Summary
Achievement-driven professional with 12+ years of experience in software development and data science. 2 years of specialized experience in Generative AI, including GPT and BERT. Accomplished Senior AVP at Wells Fargo, adept in Machine Learning and Artificial Neural Networks, with a proven track record of fostering innovation through data-driven decision-making. Honed skills in predictive modeling and quantitative analysis. Praised for exceptional teamwork and communication, significantly enhancing project outcomes without exceeding a single numerical percentage. Data-driven expert specializing in advanced statistical modeling, machine learning algorithms, and data visualization. Possesses strong problem-solving skills and deep knowledge of Python, R, SQL, and cloud-based data platforms. Dedicated to uncovering actionable insights and optimizing complex data processes.
Overview
12
12
years of professional experience
1
1
Certification
Work History
Senior AVP
Wells Fargo
06.2020 - Current
- Collaborated with business and engineering teams to promote statistical best practices in experimental design, data capture, and data analysis
- Applied data preprocessing techniques to prepare data for Machine Learning models
- Developed and implemented various regression and classification algorithms using Sklearn libraries such as Linear Regression, Decision Trees, XGBoost, and Naïve Bayes
- Created Machine Learning models for A/B testing of product content, facilitating data-driven decision-making
- Designed neural networks using TensorFlow for internal projects, including the development of automated chatbots utilizing Natural Language Processing (NLP)
- Recognized by managers, colleagues, and peers for innovation, communication, and teamwork to ensure quality, timely project completion
- Executed some internal projects using tools such as Tableau for Data Visualization
- Leveraging Generative AI (LLMs) for document summarization, chatbot development, and report generation
- Communicating technical insights to stakeholders using visualizations and dashboards (Tableau, Power BI, Streamlit)
Senior System Operation ENGINEER
Wells Fargo
05.2018 - 05.2020
Support ENGINEER
Societe Générale
08.2015 - 05.2018
Software ENGINEER
HCL Technologies
07.2013 - 08.2015
Project # 1 - Payment Operator Agent
Wells Fargo
02.2025 - 06.2025
- Built and maintained a Payments Operator Agent to automatically detect, triage, and repair failed or rejected payments within high-volume banking environments. The solution supported automated resolution of payment exceptions across domestic (ACH, RTGS) and cross-border (SWIFT, SEPA) transactions.
- Deployed a rule-based and ML hybrid system to auto-suggest fixes to operators, which were later automated into end-to-end flows using RPA bots.
- Developed automation rules and workflows to handle common payment failure cases such as: Missing or invalid beneficiary details, IBAN/BIC mismatches, Incorrect routing codes or formatting errors.
- Worked closely with compliance teams to ensure all repairs complied with KYC, AML, and sanction screening policies.
- Integrated payment repair logic using SWIFT message parsing (MT103/MT202/MT199) to extract key fields (e.g., tag 59, tag 50, tag 72) and validate against internal routing and sanction rules.
- Reduced payment repair turnaround time (TAT) by over 40%, improving SLA adherence and reducing customer complaints.
- Built dashboards for payment repair metrics (e.g., failure reasons, auto-repair rate, operator interventions) using SQL + Tableau/Power BI.
- Used machine learning and NLP to classify and cluster error descriptions from payments that failed STP (straight-through processing), leading to a 25% reduction in manual intervention.
- Increased auto-repair rate from 35% to 70%, reducing load on operations teams.
- Shortened average repair time from 20 minutes to under 5 minutes.
- Enhanced transparency through detailed audit trails and error classification, supporting internal and external audits.
Project # 2- Credit Intelligence
Wells Fargo
11.2024 - 02.2025
- Led the development of a Credit Intelligence AI system that combines traditional machine learning with Generative AI to evaluate borrower risk, generate contextual credit narratives, and automate financial document processing. The platform supported real-time decision-making for retail and SME loan underwriting.
- Fine-tuned open-source LLMs using domain-specific data to improve relevance and control over hallucinations.
- Built ML-based credit scoring models (logistic regression, XGBoost) to predict default probability based on customer financials, transactional behavior, credit history, and bureau data.
- Engineered real-time scoring and narrative generation pipelines with FastAPI + Docker, enabling instant credit assessments for digital lending journeys.
- Integrated Generative AI (LLMs like GPT-4 via OpenAI API) to: Summarize unstructured documents (e.g., income proofs, bank statements, loan agreements), Generate human-readable credit profiles and explanations for risk scores (e.g., 'Borrower shows consistent monthly income but high credit utilization'), Power a Q&A chatbot for credit analysts to query borrower insights.
- Built explainability tools using SHAP/LIME to comply with regulatory requirements and boost model transparency for underwriters.
- Collaborated with credit risk teams to incorporate policy rules and validate model outputs against real decisions.
- Reduced manual underwriting effort by 60% through GenAI-powered automation.
- Improved approval accuracy and reduced default rates by 18% compared to baseline scoring models.
- Cut document analysis turnaround time from 3 hours to under 10 minutes using LLM-powered summarization.
- Enabled real-time decision support, increasing STP (Straight Through Processing) rate for loan approvals by 40%.
- Developed a RAG pipeline using LangChain and Pinecone to fetch relevant borrower data and generate context-aware responses, ensuring accuracy and compliance.
Project # 3 - Loan Grader
Wells Fargo
04.2024 - 10.2024
- Generated a predictive model to estimate the ratings of 100,000+ customers based on their historical transactional data, including banking, card transactions, and loans.
- This model helped identify potential customers with 85% accuracy, enabling targeted marketing efforts and improving customer acquisition rates by 30%.
- Client and Team Coordination: streamlined with 5+ clients to gather detailed requirements and worked with the operations team to document and optimize 10+ manual workflows, reducing process inefficiencies by 20%.
- Effort Estimation: Reviewed 50+ work item estimates, conducted 15+ impact analyses, and participated in 10+ client meetings, improving estimation accuracy by 15%.
- Data Integration: Identified 10+ data sources and streamlined the consolidation process, reducing project preparation time by 30%.
- Data Preprocessing: Cleaned, visualized, and detected outliers in datasets exceeding 500,000 records, improving data readiness for machine learning by 25%.
- Machine Learning Evaluation: Assessed the performance of 10+ algorithms, selecting models with accuracy improvements of 10-15% for production use.
- Outcome Analysis: Analyzed results from 20+ model outputs, deriving actionable insights that contributed to 25% faster decision-making and improved business outcomes.
Project # 4 - Document Index Processing through NLP
Wells Fargo
10.2023 - 04.2024
- Operations Team manually opened and reviewed 50+ documents monthly to determine billing types, consuming 200+ hours of manual effort.
- By implementing NLP document indexing, the process of document identification was automated, reducing FTE effort by 70% and speeding up document processing by 50%.
- Process Understanding: Collaborated with the operations team to document and analyze 10+ existing workflows, identifying automation opportunities that reduced manual work by 50%.
- Client Engagement: Worked with 5+ clients to discuss and strategize the automation of document-related processes, resulting in a 30% reduction in client processing time.
- Effort Estimation: Reviewed work item estimates to improve timeline accuracy by 15%.
- Document Management: Developed a centralized platform to aggregate 100+ documents, improving access efficiency and supporting machine learning algorithms that decreased document retrieval time by 40%.
- Data Preprocessing: Conducted cleaning, visualization, and outlier detection on datasets comprising hundreds of thousands of records, enhancing machine learning input quality and minimized preprocessing time by 30%.
- Algorithm Evaluation: Measured the performance of applied algorithms and selected the most effective model.
- Model Deployment: Implemented the final model into production for operational use.
Project # 5 - Mini project on Customer behavior with change in rates
Wells Fargo
01.2023 - 03.2023
- As part of this project, prediction of behavior of a customer with the revision of rates as quick grasp for end business users.
- Effort Estimation: conducted 30+ impact analyses leading to 10% more accurate project timelines.
- Data Preprocessing: Cleaned, visualized, and eliminated outliers, reducing preprocessing time by 25% and improving model input accuracy by 15%.
- Model Evaluation and Communication: Assessed the performance of 20+ machine learning models, achieving an average accuracy improvement of 10%.
- Presented findings to 10+ business teams, enabling 30% faster decision-making based on actionable insights.
Project # 6 - Prediction of database spikes
Wells Fargo
01.2023 - 06.2023
- As part of this project, prediction of database spikes for current year based on last 5 years of data.
- Analyzed historical database performance metrics, examining 5+ years of data to identify patterns and trends leading to usage spikes exceeding 30% during peak periods.
- Designed and launched predictive machine learning models, including time series analysis and anomaly detection, achieving 90% accuracy in forecasting database activity spikes.
- Created data preprocessing pipelines that processed terabytes of raw data daily, ensuring datasets were 100% normalized and ready for predictive modelling.
- Deployed predictive models in production, integrating with real-time monitoring tools, decreased the response time to potential anomalies by 50%.
- Combined with database administrators and developers to optimize resource allocation, improving scalability by 40% based on predictive outcomes.
- Created dashboards and reports, enhancing stakeholder understanding through 10+ dynamic visualizations, increasing decision-making efficiency by 25%.
- Scheduled regular model evaluations, retraining with new data quarterly, maintaining prediction accuracy within ±5% of expected results.
- Accelerated alert systems, proactively mitigating the impact of spikes, led to downtime by 30% and maintaining 99.9% system uptime.
Education
B.Tech -
01.2010
Skills
- Data Science
- Machine Learning
- Artificial Neural Networks
- Convolutional Neural Networks
- Predictive Modelling
- Computer Vision & Image Processing
- Quantitative Analysis
- Training & Development
- Deep Learning
Certification
- Foundations of Data Science (Google): Coursera
- IBM Data Scientist: Simplilearn
- Introduction to Python Programming: Udemy
- Machine Learning, Deep Learning, NLP: Udemy
Timeline
Project # 1 - Payment Operator Agent
Wells Fargo
02.2025 - 06.2025
Project # 2- Credit Intelligence
Wells Fargo
11.2024 - 02.2025
Project # 3 - Loan Grader
Wells Fargo
04.2024 - 10.2024
Project # 4 - Document Index Processing through NLP
Wells Fargo
10.2023 - 04.2024
Project # 5 - Mini project on Customer behavior with change in rates
Wells Fargo
01.2023 - 03.2023
Project # 6 - Prediction of database spikes
Wells Fargo
01.2023 - 06.2023
Senior AVP
Wells Fargo
06.2020 - Current
Senior System Operation ENGINEER
Wells Fargo
05.2018 - 05.2020
Support ENGINEER
Societe Générale
08.2015 - 05.2018
Software ENGINEER
HCL Technologies
07.2013 - 08.2015
B.Tech -
Similar Profiles
PATRICIA M. SCHERERPATRICIA M. SCHERER
Regional Bank Private Banker II at Wells Fargo Bank / Wells Fargo AdvisorsRegional Bank Private Banker II at Wells Fargo Bank / Wells Fargo Advisors
Andrew S. HollenbachAndrew S. Hollenbach
Collateral Valuation Analysis at Wells Fargo Home Mortgage and Wells Fargo Bank NACollateral Valuation Analysis at Wells Fargo Home Mortgage and Wells Fargo Bank NA
Laxmikanth ChittampallyLaxmikanth Chittampally
Senior Operational Processor at Wells Fargo India Solution PVT Ltd, Wells FargoSenior Operational Processor at Wells Fargo India Solution PVT Ltd, Wells Fargo
Mayank OberoiMayank Oberoi
Financial Accounting Associate/ Alteryx SME at Wells Fargo International Solutions Private LTD (Wells Fargo)Financial Accounting Associate/ Alteryx SME at Wells Fargo International Solutions Private LTD (Wells Fargo)
Ayushi SharmaAyushi Sharma
Emerging Architect at Shell India Markets Pvt LtdEmerging Architect at Shell India Markets Pvt Ltd