Souvik Ghosh

Staff AI/ML & Generative AI Architect

Agentic AI & AI Platforms | AWS Cloud | Data & Platform Architecture

|

Souvik Ghosh

About Me

Staff Platform Engineer & Solutions Architect with nearly 16 years of experience architecting and delivering enterprise-grade full-stack, cloud-native, data, and AI platforms, including 5+ years of production Generative AI implementation.

Spearheaded Agentic AI, Multi-Agent, RAG, and LLM architectures leveraging AWS Bedrock, LangGraph, LangChain, OpenAI, Claude, and hybrid search, translating complex enterprise challenges into scalable intelligent solutions. Engineered large-scale data ecosystems using Python, Spark, Airflow, Kafka, AWS Glue, and Athena, processing cloud events daily across multiple AWS accounts to power analytics, automation, and AI-driven decision intelligence.

Recognized for end-to-end architecture ownership, customer-embedded engineering, executive stakeholder engagement, and reusable AI/cloud accelerators, consistently bridging business strategy, technology architecture, and production execution across complex enterprise environments.

Key Career Highlights

  • Enterprise GenAI & Agentic AI: Architected enterprise-grade GenAI, Agentic AI, RAG and multi-agent platforms delivering 75% lower manual effort and 40% faster turnaround.
  • Large-Scale Data & Telemetry: Architected data and operational platforms processing 15M+ cloud events daily across 1,700+ AWS accounts with Spark, Airflow, Kafka, Glue, EMR, Athena and GenAI.
  • Operational Reliability & LLMOps: Established LLMOps/MLOps, evaluation, monitoring and reliability patterns contributing to 99.9% availability, 30% MTTR reduction, and 20% lower cloud costs.
  • Reusable Accelerators: Built reusable architecture patterns and AI/cloud accelerators reducing new AI use-case build effort by ~35%.

Work Experience

Staff Platform Engineer - Cloud & Gen AI

Commonwealth Bank of Australia (CBA) India, Bengaluru | Jun 2025 - Present

  • Architected an AI-powered operational intelligence platform on AWS, spanning data ingestion, transformation, orchestration, model integration, backend services and UX—automating 10+ enterprise workflows and reducing manual investigation effort by ~40%.
  • Designed large-scale telemetry and data-processing architecture using Apache Spark and Airflow, processing 15M+ CloudTrail and AWS Config events daily across 1,700+ AWS accounts, enabling enterprise-wide analytics, contextual reasoning and AI-assisted operational decision-making.
  • Designed AI-assisted incident investigation and runbook orchestration capabilities, combining infrastructure telemetry, GenAI reasoning and workflow automation to reduce MTTR by 30% and improve troubleshooting consistency by 50%.
  • Established reusable AI platform patterns and reference architectures incorporating Temporal Workflow-as-Code, Terraform and CloudFormation, reducing new AI use-case development effort by ~35% and accelerating repeatable enterprise adoption.
  • Drove iterative production optimization across Platform Engineering, SRE, Product & executive stakeholders, improving AI recommendation accuracy by ~25% across successive releases while aligning architecture with reliability, scalability and operational requirements.
  • Bridged AI engineering, cloud platform and application architecture, translating operational problems into production-ready solutions while balancing accuracy, latency, reliability, security, cost and developer experience.

Data & AI Engineer | GenAI Solution Architecture & Customer-Embedded Delivery

Capgemini Engineering, Bengaluru | May 2022 - Jun 2025

  • Led enterprise discovery and solution architecture with engineering teams, business stakeholders and C-suite sponsors, translating ambiguous business problems into production-grade AI, data and full-stack solutions that reduced manual effort by 75% and accelerated decision-making by 40%.
  • Architected end-to-end AI applications spanning Python/FastAPI services, TypeScript/React applications, Spark data pipelines, curated data layers, APIs and deployment infrastructure, taking solutions from discovery through production and user adoption.
  • Designed and productionized Agentic RAG architectures using LangGraph, integrating multi-step reasoning, semantic retrieval, tool orchestration and contextual grounding, improving answer accuracy by 30% and reducing user resolution time by 45%.
  • Productized reusable LLM application patterns using LangChain, LangGraph, LangSmith and Amazon Bedrock, accelerating feature delivery by 50% across 8+ enterprise use cases.
  • Established GenAI evaluation and reliability practices using LangSmith and RAGAS, combining prompt optimization, structured outputs and function calling to improve response consistency by 35%, reduce hallucinations by 25%, and cut manual validation effort by 60%.
  • Architected enterprise tool and API integration patterns, connecting 15+ business APIs and enterprise systems to AI agents to enable autonomous workflow execution and eliminate ~50% of repetitive manual activity.
  • Partnered with engineering leadership, product teams and client stakeholders on solution envisioning, architecture decisions and GenAI capability building, contributing to reusable accelerators and client proposals.

Data Science Practitioner

Tata Consultancy Services (TCS) | Mar 2020 - May 2022 | Kolkata, India

  • Delivered production-grade ML/DL solutions across predictive analytics, NLP, anomaly detection & computer vision, including a CNN-based corrosion/damage detection solution that reduced manual inspection effort by 70% & generated approximately AUD 750K in annual savings.
  • Performed large-scale EDA across millions of application logs and automated KPI intelligence dashboards, driving 20% higher user engagement, 15% faster load times, and 60% less manual reporting.

Cloud Applications Solutions Architect

Tata Consultancy Services (TCS) | Jun 2018 - May 2022 | Kolkata, India

  • Architected secure, scalable AWS cloud-native applications using Node.js/TypeScript, Cognito, IAM, S3, SES, SNS, Lambda, and DynamoDB Streams, aligning architecture with enterprise NFRs.
  • Modernized API ecosystems using AWS Lambda & API Gateway, delivering 40% faster API response times & 30% lower infrastructure costs.
  • Industrialized CI/CD with AWS CodeBuild & CodePipeline, reducing deployment cycles by 70% & accelerating automated defect detection.

Cloud Applications Developer

Tata Consultancy Services (TCS) | May 2016 - May 2018 | Kolkata, India

  • Engineered 5+ cloud-native full-stack applications on AWS serverless architecture using Node.js/TypeScript, reducing infrastructure overhead by 30% and accelerating release cycles by 40%.
  • Orchestrated cross-functional delivery across Product, QA, Infrastructure, improving requirement traceability & reducing delivery gaps by 25%.

Onsite Project Coordinator – Client-Facing

Tata Consultancy Services (TCS) | May 2015 - May 2016 | London, UK

  • Led onsite technical delivery and cross-functional coordination for a UK enterprise client, improving execution efficiency by 30%.
  • Facilitated 20+ stakeholder workshops and institutionalized decision governance, reducing communication gaps by 40% and achieving 100% on-time delivery.

Application Developer & Team Lead

Tata Consultancy Services (TCS) | Jan 2011 - Apr 2015 | Chennai, India

  • Engineered banking application modules and API integrations with legacy core systems, reducing deployment timelines by 35% while sustaining 99.5% application availability.
  • Led and mentored a 6-member engineering team across the SDLC, strengthening engineering quality and maintaining zero production-critical incidents for 18 months.

Core Competencies & IT Skills

AI & Generative AI

Agentic AI Multi-Agent Systems RAG & Hybrid Search LLM Application Architecture LangGraph LangChain LangSmith LlamaIndex Amazon Bedrock OpenAI SDK Anthropic Claude Azure OpenAI Prompt Engineering Structured Outputs Tool/Function Calling Embeddings RAGAS LLM Evaluation LLMOps

AI Platform & Agent Architecture

Agent Runtime & Orchestration Tool Integration Context Engineering Memory Patterns Guardrails Workflow Orchestration AI Automation Intelligent Workflows Agent Evaluation AI Observability Model/Prompt Optimization

Cloud & Platform Architecture

AWS AWS Lambda ECS / EKS Amazon S3 DynamoDB RDS API Gateway Cognito IAM Step Functions Bedrock Terraform CloudFormation Docker Kubernetes CI/CD

Data Engineering & Big Data

Python SQL PySpark Apache Spark Apache Airflow Kafka AWS Glue EMR Athena Batch & Streaming Architecture Data Lake Architecture Data Governance & Cataloguing

Search & Knowledge Engineering

OpenSearch Pinecone ChromaDB FAISS Vector Search Hybrid Search Embedding Engineering Knowledge Retrieval

Application & Integration Architecture

TypeScript JavaScript Node.js React FastAPI REST / GraphQL APIs Microservices Event-Driven Architecture API Integration Distributed Systems

Leadership & Delivery

Solution Architecture Architecture Governance Technical Strategy Customer-Embedded Engineering Executive Stakeholder Management Presales & Solution Envisioning Reusable Accelerators Cross-Functional Technical Leadership

Education

Advanced Certificate Program in Generative AI

upGrad, India

2024

Certified

Master of Technology - M.Tech (Data Science & Engineering)

Birla Institute of Technology & Science (BITS), Pilani, India

2022

Grade: CGPA - 8.74/10

Bachelor of Engineering - B.E. (Computer Science & Engineering)

Visvesvaraya Technological University (VTU), Bengaluru, India

2010

Grade: First Class

Publications

Democratization of NOSQL Document-Database over Relational Database - Comparative Case Study: Cloud Kitchen

International Research Journal of Engineering and Technology (IRJET)

Apr 2022

"The online food delivery market revenue is expected to reach $140 Million+ worldwide. As a result, the new model of dining has already taken a paradigm shift to capture this extravagant market. Consumers are now shifting from dining to delivery/take-away model. Here takes birth the concept of Cloud Kitchen – orders placed, meals cooked, packed and then dispatched immediately to the respective delivery locations by the assigned fleet, E.g. Swiggy, Zomato, UberEats. Apart from Lower Real Estate costs, saving overhead costs, convenience, etc., one should also consider what it takes to setup a proper datastore infrastructure to cater to the need of these cloud kitchens. With respect to that need, we intend to perform a comparative analysis on database(s) (NOSQL Vs. SQL) one needs to choose for faster data access, order placement, bill generation, delivery notification settings, standard feedback loops and customer sentiment analysis for better service. We have done a detailed analysis based on expected / possible database CRUD operations that could get performed throughout one order cycle – right from placement till delivery and post-delivery customer feedbacks and have given our proposal of database choice based on that mode and complexity of operation."

Certifications

AWS Certified Generative AI Developer – Professional (Early Adopter)

Amazon Web Services (AWS) | 2026

View Credential

HashiCorp Certified: Terraform Associate (004)

HashiCorp | 2026

View Credential

AWS Certified Machine Learning – Specialty

Amazon Web Services (AWS) | 2026

View Credential

AWS Certified Machine Learning Engineer – Associate

Amazon Web Services (AWS) | 2026

View Credential

AWS Certified AI Practitioner

Amazon Web Services (AWS) | 2025

View Credential

AWS Certified Solutions Architect – Associate

Amazon Web Services (AWS) | 2020

View Credential

Model Context Protocol: Advanced Topics

Anthropic | 2025

View Credential

Advanced Certificate in Generative AI

upGrad | 2024

View Credential

Oracle Cloud Infrastructure 2024 Generative AI Certified Professional

Oracle | 2024

View Credential

Microsoft Certified: Azure Data Scientist Associate | DP -100

Microsoft | 2022

View Credential

Microsoft Certified: Azure AI Fundamentals

Microsoft | 2022

View Credential

Microsoft Certified Azure Fundamentals

Microsoft | 2022

View Credential

Badge Wallet

Projects

Versatile Multimodal & Multilingual AI for Task Automation

Capgemini Engineering Client Project

Aug 2024 - Mar 2025

Developed a versatile multimodal and multilingual AI product for end-to-end customer support.

Prompt2plot - AI-driven system for data visualizations

Capgemini Engineeing Client Project

Aug 2022 - Nov 2022

Developed Prompt2plot, an AI-driven system that autonomously generates diverse data visualizations.

AWS cloud-based data lake for real-time analytics

Capgemini Engineering Client Project

May 2022 - Nov 2022

Designed and implemented an AWS cloud-based data lake enabling real-time analytics.

Recognitions

"Care"

Commonwealth Bank

Sep 2025

Appreciation for the support provided in GenAI activities and DataLake enhancement.

"Courage"

Commonwealth Bank

Nov 2025

For proactively onboarding Project Coral and successfully leading the team to deliver the project on time.

"Commitment"

Commonwealth Bank

Nov 2025

Recognition for my invovement in Datafest, my contribution to the Resilience space and my building copilot agent for On-call Support Activities.

"Wow Award Q4 2023"

Capgemini Engineering

Mar 2024

Appreciation for the support provided in GenAI activities and collaboration on several proposals that resulted in successful submission to the client

"Wow Award Q3 2023"

Capgemini Engineering

Dec 2023

Recognition for contributions to competency building in Generative AI and assistance in pre-sales activities.

"Star of the Month Award"

Tata Consultancy Services

Jun 2014

Recognition for handling Severity-1 incident efficiently and restoring the service without any major impact on the business.

Many More...

And counting!

Testimonials

Kaushik Ganguly

Senior Manager - Data & AI | Principal Architect - Agentic AI Enterprise & Solution Architecture

Accenture

"Not often would you come across a talent like Souvik. His ability to seamlessly transition to cloud technologies has not only transformed our team’s operations but also set a benchmark for excellence within the industry. Souvik's commitment to continuous learning and growth is evident as he currently excels in the AI/ML field, where he is making significant contributions and driving innovative solutions. Beyond his technical skills, Souvik is an exceptional team member. His collaborative spirit and positive attitude create an environment where everyone feels valued and motivated. He is someone you can always count on, whether it's for support on a challenging project or sharing valuable insights during team discussions. Souvik's analytical mindset and problem-solving abilities are impressive, making him a key asset to any organization. He consistently demonstrates initiative and creativity, ensuring that he not only meets but exceeds expectations. I have no doubt that he will continue to achieve great things and inspire those around him. In short, I wholeheartedly recommend Souvik to any organization looking for a dedicated and talented professional. His expertise, work ethic, and collaborative nature will undoubtedly make him a standout contributor."

Bharat Kumar Behera

Cloud Consultant

Tata Consultancy Services

"I am pleased to recommend Souvik, with whom I have had the opportunity to work closely at our organization. Souvik has consistently demonstrated strong expertise in Generative AI, including hands-on experience with large language models, prompt engineering, model integration, and real-world AI use cases. He has a solid understanding of both the theoretical and practical aspects of GenAI systems and is quick to translate ideas into scalable, working solutions. What truly sets Souvik apart is his ability to combine technical depth with problem-solving mindset. He approaches complex challenges methodically, stays up to date with the rapidly evolving GenAI ecosystem, and contributes meaningfully to discussions around architecture, optimization, and responsible AI usage. His work reflects clarity, innovation, and attention to detail. Beyond technical skills, Souvik is a dependable team player. He communicates clearly, collaborates well across teams, and takes ownership of his responsibilities. He has shown maturity, adaptability, and a strong willingness to learn — qualities that are essential in fast-moving AI-driven environments. I am confident that Souvik will be a valuable asset to any team working in the field of Generative AI. "

Krunal Soni

Data Analyst

Worldpanel

"I've had the privilege of working alongside Souvik Ghosh, and his depth in AI/ML and Generative AI is truly exceptional. With a sharp mind for building scalable, production-ready systems, Souvik Sir bridges deep technical know-how with real-world business impact... His ability to adapt across domains and lead with clarity makes him a standout in the AI space. Any team would be lucky to have him driving their next AI initiative."

Akshay Kumar

Senior Software Engineer

AirAsia MOVE

"Working with Souvik has been a masterclass in applied AI innovation. His rare blend of deep technical expertise and strategic thinking makes him an exceptional AI/ML and Generative AI specialist. Whether it's building LLM-powered solutions, optimizing inference pipelines, or leading enterprise AI transformations, Souvik delivers scalable impact across domains. His passion for Agentic AI and forward-thinking approach keeps him ahead of the curve. A true collaborator and thought leader, Souvik is someone you want on your team when tackling complex, high-stakes AI challenges."

Sejaal Pattaddia

PTA | Delivery Operations Manager

Microsoft

"I’ve had the opportunity to collaborate with Souvik on AI-driven initiatives, and his technical depth in ML, NLP, and generative AI is outstanding. He has a strong command of Python, TensorFlow, and PyTorch and is highly proficient in building and deploying scalable models across AWS, Azure, and GCP. One of his strengths is translating business needs into robust, production-ready AI solutions—from fine-tuning LLMs for specific use cases to optimizing inference pipelines and deploying them securely at scale. His recent work in agentic AI is a great example of pushing boundaries with autonomous, goal-oriented agents. If you're looking for someone who can architect enterprise-grade AI systems with performance, scalability, and impact in mind, Souvik is the engineer you want on your team."

Hari Gupta

Sales Manager

Touracle

"Souvik is a standout AI/ML expert with deep hands-on experience in GenAI, LLMs, and scalable solutions. He blends technical excellence with business acumen, delivering impactful, enterprise-ready innovations across domains. A true asset to any forward-thinking team."

Pooja Patil

Team Lead Infrastructure Management

Nokia

"Souvik is one of the most visionary and technically accomplished professionals I’ve had the pleasure to work with. With a rare blend of deep AI/ML expertise and real-world domain understanding, he consistently delivers enterprise-grade solutions that are both innovative and impactful. His ability to design scalable GenAI systems, optimize LLMs, and lead full ML pipelines—from ideation to deployment—is truly impressive. Souvik's passion for pushing boundaries, especially in the Agentic AI space, shows his forward-thinking mindset. He’s not just solving today’s problems—he’s building the future. Any team would be lucky to have his leadership, insight, and technical excellence."

Prasanna Lakshmi Karredla

Lead, Services Application Developer

Schneider Electric

" have had the pleasure of working with Souvik, an exceptional AI-ML & Generative AI Specialist. His expertise spans across model training, model fine-tuning, LLM integration, prompt engineering, RAG implementation, and scalability optimization. Souvik's deep understanding of these domains is truly impressive. Souvik consistently demonstrates a remarkable ability to tackle complex problems with innovative solutions. His proficiency in model training and fine-tuning ensures that our AI models perform at their best. His skills in LLM integration and prompt engineering have significantly enhanced our projects, making them more efficient and effective. Moreover, Souvik's expertise in RAG implementation and scalability optimization has been invaluable in ensuring our systems are robust and scalable. His dedication to staying updated with the latest advancements in AI and ML is evident in his work, and he is always willing to share his knowledge with the team. Souvik is not only a technical expert but also a great collaborator. He communicates complex ideas clearly and works seamlessly with others to achieve common goals. His positive attitude and willingness to help make him a pleasure to work with. I highly recommend Souvik for any role that requires deep expertise in AI and ML. He is a true asset to any team."

Mohd Rohil

Product manager

Cendyn

"I’ve had the pleasure of working with Souvik Ghosh, and I can confidently say he’s one of the most dependable and forward-thinking AI specialists I’ve encountered. Whether it’s training custom models, fine-tuning LLMs, or optimizing large-scale AI systems, Souvik brings both deep technical expertise and a clear, practical mindset to the table. What truly sets him apart is his ability to break down complex AI/ML concepts into simple, actionable steps for the team. He’s not just a brilliant technologist—he’s also a fantastic collaborator. From RAG implementations to advanced prompt engineering, I’ve seen Souvik deliver outcomes that are not only technically sound but also aligned with business needs. If you’re looking for someone who can seamlessly connect the dots between cutting-edge AI research and real-world impact, Souvik is that person. I’d work with him again in a heartbeat."

Hobbies & Interests

Photography

Capturing moments and telling stories through the lens.

Photography 1 Photography 2 Photography 3 Photography 4

Travelling

Discovering new cultures, cuisines, and landscapes around the world.

Travel 1 Travel 2 Travel 3 Travel 4

Let's Connect

Bengaluru, Karnataka, India

Talk to My Resume
Hi! I'm Souvik's AI assistant. Ask me anything about his experience, skills, or projects! 🤖