Neeladri Shekhar Pal

Neeladri Shekhar Pal

He / Him / His

Lead Data Scientist · Generative AI & ML · Forecasting & LLMOps

Kolkata, India

About

Over the past 6+ years, I have led production AI/ML initiatives for Insurance, Petrochemicals, and Financial Services clients across the US, Australia, and the Middle East — translating ambiguous problems into measurable impact.

My recent focus is Generative AI, Agentic AI (LangGraph), and time-series forecasting. I have shipped a conversational claims agent operating at ~99% success across 22,000+ production queries, and a multi-agent fraud detection platform that cut claim adjudication from hours to under 2 minutes. Earlier engagements delivered >95% forecasting accuracy and ~35% process optimization for global clients.

Engineer by education, data whisperer by profession — a firm believer that numbers tell captivating stories. Outside of the data universe, I'm all about finding balance.

Experience

Lead Assistant Manager — Data Science
EXL Service Limited · Nov 2023 – Present
Client: One of the largest Life Insurance businesses in the USA
WEA Claim Analytics Agent
LangGraph · Python · Pandas · Flask · Plotly Dash
  • Conversational AI for Claims Analytics: Architected and deployed a production LangGraph-based agent on Cortex Lab that translates natural-language questions into executable code, querying 7M+ claim activity records from MIMICA in real time.
  • Multi-Stage Reasoning Pipeline: LLM-driven workflow with intent triage (Conversational / Simple-Data / Multi-Step), dynamic plan decomposition, code generation, sandboxed execution, correctness verification, and a self-correction loop with up to 5 retries and persistent error memory.
  • Code Safety & Guardrails: 6-layer safety framework covering prompt-injection detection (instruction override, persona hijack, prompt extraction, delimiter injection, code injection, PII exfiltration), sandboxed execution with import blocking, timeouts, output caps, schema/answer caching, and full observability via S3-persisted trace logs.
  • Production Deployment: Gunicorn + Flask + Plotly Dash with async polling-based streaming for real-time response delivery.
  • Outcome: ~99% success rate across 22,000+ production queries over 150 active days (100% in the most recent week); enabled stakeholders to surface SME performance, tool usage, and claim-duration insights without writing code.
ClaimIQ Analytics Accelerator — Multi-Agent Fraud Detection (EXL IP: Accelerate.AI)
LangGraph · FastAPI · React · Docker · WebSocket
  • Multi-Agent Architecture: LangGraph-orchestrated pipeline of 7 specialized agents handling FNOL intake, loss-cause classification, and parallel fraud adjudication across structured, document, visual, and network-graph modalities.
  • Hybrid ML + LLM Reasoning: Combined a pre-trained XGBoost classifier with LLM-powered agents through a provider-agnostic layer supporting Azure AI Foundry and Amazon Bedrock for flexible model switching.
  • Visual & Document Forensics: Image authenticity checks (reverse search, AI-generated detection) and cross-document discrepancy reasoning to surface manipulated evidence and inconsistent claim narratives.
  • Network Graph Intelligence: Fraud-ring detection over historical claims using entity-level (garage, broker, customer, ZIP) overlap analysis, visualized as an interactive graph for adjusters.
  • Smart Triage & PII Compliance: Auto-routed claims to Auto-Approve, Manual Review, or SIU queues with audit-logged decisions; upstream PII masking via Microsoft Presidio for GDPR/HIPAA-aligned data handling.
  • Outcome: Reduced claim adjudication from hours to under 2 minutes with explainable, evidence-backed verdicts that adjusters can defend in SIU referrals.
Non-Catastrophic Claims Forecasting
Python · TensorFlow · Plotly
  • Meta-Modeling Architecture: Multi-layer forecasting framework spanning 48 U.S. states and 7 lines of business, combining Chronos LLM, TabPFN Timeseries, TimesFM, PatchTST, NBEATS, and Ridge regression on residuals for long-horizon prediction.
  • Accuracy & Stability: >95% average accuracy on a 56-month validation window using a rolling 26-week horizon at weekly granularity — consistent across market and seasonal regimes.
  • Outcome: Enabled state-level dynamic resource allocation and capacity planning across regions and lines of business.
Insurance Policy Evaluation Optimization
Python · SQL · Excel
  • Questionnaire Optimization: Analyzed 107 policy evaluation parameters using Cramér's V and Phi-coefficient association tests to identify redundant and interdependent variables.
  • Feature Prioritization: Used 2-way / 3-way ANOVA against claim outcomes (high SSR vs. SSE) to retain categorical questions with high discriminative power.
  • Outcome: Reduced the questionnaire to 60–70 items per state-specific compliance rules — streamlined policy evaluation by ~35% and shortened the underwriting lifecycle by ~14%.
Apprentice Leader — Data Science Lead
Mu-Sigma Business Solutions Pvt. Ltd. · May 2021 – Nov 2023
Clients: Largest petrochemical producer in Saudi Arabia · Largest life insurer in Australia
Intermittent Demand Forecasting via Meta-Modeling (S&OP)
Python · RapidMiner · Mar 2023 – Nov 2023
  • Cannibalization Detection: Identified product cannibalism and upgrade transitions using cointegration and distributional tests across the SKU portfolio.
  • Product Association: Discovered SKU interconnections and built alias mappings from historical purchase patterns using Market Basket Analysis.
  • Modeling: Designed a meta-modeling solution combining Linear Dynamical Systems and Deep Auto-Regressive Recurrent Networks, scaled to ~2,600 SKUs across ~49 countries in the MEAF region.
  • Leadership: Led the data science squad, owned client communications, and drove the transition to an Agile-Waterfall hybrid — improving utilization tracking, velocity estimation, and delivery quality.
Sales Price Prediction — Polyethylene Sub-products
Python · RapidMiner
  • Multivariate Time Series: Deployed SARIMAX models tailored to each sub-product across global sales offices.
  • Outcome: ~97% accuracy across 22 sales offices globally for all 3 polyethylene sub-products.
Document Classification & Automation
Python · SQL
  • Performed EDA and pre-processing to build training data from thousands of customer emails.
  • Classified claims documents using XGBoost on text from emails, lodgment files, and audio transcripts.
  • Built a model-monitoring dashboard in Plotly Dash to track target drift and feature drift in production.
Founder
Quantbot Securities Pvt. Ltd. · Jun 2018 – Feb 2026
Founded and scaled an end-to-end algorithmic trading platform serving 9,000+ users across India for quantitative research and systematic trading on Indian and global equity markets.
  • Platform: Microservices stack with Python backend on Kubernetes, Redis for caching and order queueing, GitHub Actions for CI/CD, and a React frontend for strategy visualization and live trading.
  • Quant Research: Processed multi-year high-frequency equity data to surface tradable signals; tested mean-reversion and momentum hypotheses using statistical tests (ADF, autocorrelation, t-test, Chi-Square) for robustness.
  • Leadership: Owned product strategy, architecture, hiring, and client outreach — scaling from concept to 9K+ active users in production.
Secretary — Insight, Data Science Society
IMI New Delhi · Mar 2020 – Mar 2021
Led society administration and authored tutorial blogs on machine learning fundamentals — including K-NN and Market Basket Analysis — to broaden access to applied data science within the campus.

Core Competencies

Generative AI & LLMs
LangGraph LangChain Multi-Agent Systems LLM Fine-Tuning (LoRA, DAPT) RAG Prompt Engineering Safety Guardrails Hugging Face Transformers FAISS ChromaDB HNSW Knowledge Graphs
Machine Learning
Chronos TimesFM TabPFN PatchTST NBEATS SARIMAX FbProphet Meta-Modeling & Ensembles XGBoost Random Forest Gradient Boosting SVM RNN / LSTM Auto-Regressive Networks
MLOps & Engineering
Docker FastAPI Flask Gunicorn Redis CI/CD On-Prem LLMOps Drift & Stability Monitoring Microservices WebSocket Streaming
Programming & Analytics
Python R SQL TensorFlow PyTorch Plotly Dash Power BI Tableau
Leadership & Domain
Team Leadership Stakeholder Management Project Estimation Agile-Waterfall Hybrid Mentoring Insurance (Claims, Underwriting, Fraud) Petrochemicals (S&OP) BFSI

Education

PGDM (MBA), Finance & Decision Science
International Management Institute (IMI), New Delhi · 2019 – 2021 · 57.60%
Specialized in Finance and Decision Science, with focus on data science applications in business contexts.
B. Tech, Construction Engineering
Jadavpur University, Kolkata · 2014 – 2018 · 65.45%
Engineering foundation with analytical and problem-solving skills that translate well to data science applications.
Higher Secondary (12th), WBCHSE
Phanindra Dev Institution, Jalpaiguri · 2014 · 90.60%
Secondary (10th), WBBSE
Phanindra Dev Institution, Jalpaiguri · 2012 · 91.43%

Featured Work & Publications

Market Basket Analysis in Management Research (using R)
Insight - Data Science Society, IMI Delhi
Comprehensive analysis of market basket analysis applications in business intelligence environment, helping retailers understand customer purchasing habits and anticipate behaviors.
→ Read Full Article
K-Nearest Neighbours (K-NN) Algorithm from Scratch (in R)
Insight - Data Science Society, IMI Delhi
Detailed tutorial explaining K-NN algorithm implementation from scratch with practical examples, covering the theoretical foundations and hands-on applications of this fundamental machine learning algorithm.
→ Read Full Article

Certifications

Fundamentals of Reinforcement Learning — University of Alberta (Coursera) Neural Networks & Deep Learning — DeepLearning.AI Analyzing & Visualizing Data with Power BI — Microsoft Introduction to Time Series Forecasting in R — Udemy

Awards & Recognition

Beyond Work

Contact

santu96415@gmail.com +91 70035 74590 LinkedIn