Pranjal Shukla

Data Science Manager | AI & Cloud Architect

Building intelligent systems that drive business impact

Pranjal Shukla

About Me

Data Science Manager and Certified AWS & Databricks Professional with 5+ years of experience in AI/ML and cloud architecture. Currently leading a talented team of data scientists at Shriram Finance, where I drive digital innovation through intelligent systems and predictive modeling.

My expertise lies in architecting Agentic RAG platforms, LLM fine-tuning pipelines, and serverless GenAI infrastructures that bridge the gap between complex data science and scalable, business-critical production.

₹298Cr

Monthly Conversions

60M+

SMS Processed

75%

Cost Reduction

80%

Risk Capture Rate

Experience

Dec 2024 - Present

Data Science Manager

Shriram Finance | Bangalore, India

  • Leading a 7-member team in AI & Data Science division
  • Built Agentic RAG platform: 15% FD conversion uplift
  • LLM fine-tuning engine processing 60M+ SMS
AWS Bedrock LangChain Gemma 3 CatBoost
Sep 2023 - Dec 2024

Data Solution Architect

Latent Plane | London, UK

  • Engineered async infrastructure: 95% resolution, <2s latency
  • Reduced annual support costs by ~75%
  • Claude 3 orchestration for intent analysis
AWS Lambda Claude 3 MongoDB SES
May 2023 - Sep 2023

Product Manager Intern

Datalink Software | Tampa, Florida

  • Databricks ETL for 10,000+ patient records
  • Delivered 12 client features, 25% efficiency gain
Databricks ETL Product Management
June 2020 - Aug 2022

Technical Associate

Intellexs Investment Solutions | Noida, India

  • Cloud migration to AWS RDS with serverless ETL
  • Credit risk model: 85% accuracy on 30K portfolio
AWS RDS DMS Kinesis XGBoost

Signature Projects

Agentic RAG Platform

Shriram Finance

+15% Conversion

Multi-tool Fixed Deposit Agent using AWS Bedrock and Strands SDK with real-time financial data retrieval. Built semantic routing framework for complex query resolution.

AWS Bedrock Strands SDK Crawl4AI
₹298Cr/mo

LLM Fine-Tuning Engine

Shriram Finance

60M+ SMS

End-to-end SMS Engine using Gemma 3 27B fine-tuned with Unsloth. Multi-stage pipeline from noise removal to cohort generation, identifying 50+ financial and behavioral segments.

Gemma 3 27B Unsloth MoEngage
50+ Segments

Predictive Collection Model

Shriram Finance

80% Capture Rate

CatBoost-based bounce prediction classification model with 70+ engineered features. Enabled prioritized collection strategies and significantly reduced EMI bounce impact.

CatBoost Feature Engineering SQL
₹70Cr/mo

Serverless GenAI Architecture

Latent Plane

75% Cost Reduction

Highly scalable async infrastructure using AWS Lambda, API Gateway. Intelligent orchestration with Claude 3 for real-time intent and sentiment analysis with 6-year compliance lifecycle.

Claude 3 AWS Lambda MongoDB
<2s latency

Skills & Certifications

Technical Skills

AI & Machine Learning

Machine Learning Deep Learning NLP PyTorch XGBoost CatBoost scikit-learn

Generative AI & LLMs

LangChain RAG AWS Bedrock Claude OpenSearch Pinecone FAISS

Cloud & Infrastructure

AWS Lambda API Gateway SageMaker Databricks FastAPI

Leadership

Team Management Stakeholder Relations Product Strategy

Certifications

AWS

AWS Certified Solutions Architect

Associate

AWS

AWS Certified Cloud Practitioner

Foundational

DATA

Databricks Certified Data Engineer

Professional

Education

Master's in Business Analytics

AI & Cloud Specialization

George Washington University | 2022-2023

B.Tech in Computer Science

Vellore Institute of Technology | 2016-2020

Let's Connect

Open to discussing AI/ML opportunities, consulting projects, or just having a conversation about technology.

Bangalore, India | Delhi NCR, India