Titas Biswas

Titas Biswas

Technical Architect

📍 Bangalore, India

Resume

// distributed systems · enterprise architecture · AI

Building Distributed
Systems That Think

Staff-level architect bridging enterprise distributed systems and AI architecture. From petabyte-scale data pipelines to production RAG — shipping systems that matter.

16+ Years of Experience
11+ Part AI Series
5+ Production Projects

Built systems at

Tesco VMware Bazaarvoice IBM
Ask me anything about my work, projects, or experience

AI: Through an Architect's Lens

Exploring AI/ML concepts from a distributed systems architect's perspective — bridging theory with production-grade engineering.

5/5
0B conceptual

Part 0B: From Sequences to Transformers

The journey from “words in order” to “understanding meaning” — how we taught machines to process language, and why the Transformer changed everything.

0A conceptual

Part 0A: Neural Networks & The Learning Mechanism

Building deep intuition from first principles — how neural networks actually learn, explained step by step for engineers who want to truly understand the machinery before architecting with it.

1A conceptual

Part 1A: Understanding the LLM Machine

The technical foundations that drive every architectural decision — transformers, embeddings, and tokenization- are explained through the lens of cost, performance, and trade-offs.

1B conceptual

Part 1B: Making Decisions with LLMs

From model selection to production reliability — the decision frameworks that separate prototype AI from enterprise systems.

2A conceptual

Part 2A: Production RAG: What Tutorials Don’t Teach You

From naive retrieval to production-grade systems — the architectural patterns, chunking strategies, and retrieval engineering that separate demo RAG from enterprise RAG.

Season 1: Kubernetes: The Container Odyssey

This is a 12-part Kubernetes tutorial series written for experienced software engineers who are new to Kubernetes. It follows the story of Alex, a senior backend engineer with a decade of experience in monoliths and VMs, who joins a fast-growing startup called NovaCraft where everything runs on Kubernetes. Each chapter solves a real problem Alex faces at NovaCraft, combining deep conceptual explanations with practical, runnable examples that you can follow on your macOS laptop. The series progresses from foundational concepts to production-ready practices, building on each previous chapter.

13/13
0 conceptual

The Container Odyssey: A Kubernetes Tutorial Series: Season 1

This is a 12-part Kubernetes tutorial series written for **experienced software engineers** who are new to Kubernetes. It follows the story of Alex, a senior backend engineer with a decade of experience in monoliths and VMs, who joins a fast-growing startup called NovaCraft where everything runs on Kubernetes. Each chapter solves a real problem Alex faces at NovaCraft, combining deep conceptual explanations with practical, runnable examples that you can follow on your macOS laptop. The series progresses from foundational concepts to production-ready practices, building on each previous chapter.

1 conceptual

Part 1: "The New Gig" — Why Kubernetes Exists and What Problem It Solves

"The New Gig" — Why Kubernetes Exists and What Problem It Solves

2 conceptual

Part 2: "Containers 101" — Docker Fundamentals You Actually Need for K8s

"Containers 101" — Docker Fundamentals You Actually Need for K8s

3 conceptual

Part 3: The First Deploy — Pods, the Atomic Unit of Kubernetes

The First Deploy — Pods, the Atomic Unit of Kubernetes

4 conceptual

Part 4: "Scaling the Team" — ReplicaSets, Deployments, and Rolling Updates

"Scaling the Team" — ReplicaSets, Deployments, and Rolling Updates

5 conceptual

Part 5: "Opening the Doors" — Services, Networking, and Ingress

"Opening the Doors" — Services, Networking, and Ingress

6 conceptual

Part 6: "The Config Puzzle" — ConfigMaps, Secrets, and Environment Management

"The Config Puzzle" — ConfigMaps, Secrets, and Environment Management

7 conceptual

Part 7: "Persistent Memories" — Storage, Volumes, and StatefulSets

"Persistent Memories" — Storage, Volumes, and StatefulSets

8 conceptual

Part 8: The Night Watch — Health Checks, Resource Management, and Observability

The Night Watch — Health Checks, Resource Management, and Observability

9 conceptual

Part 9: "The Assembly Line" — Jobs, CronJobs, and DaemonSets

"The Assembly Line" — Jobs, CronJobs, and DaemonSets

10 conceptual

Part 10: "Fortress K8s" — RBAC, Network Policies, and Security Best Practices

"Fortress K8s" — RBAC, Network Policies, and Security Best Practices

11 conceptual

Part 11: "The Helm of the Ship" — Helm, Kustomize, and Templating

Part 11: "The Helm of the Ship" — Helm, Kustomize, and Templating

12 conceptual

Part 12: "Production Ready" — CI/CD, GitOps, and the Road Ahead

"Production Ready" — CI/CD, GitOps, and the Road Ahead

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time.

A python series designed for absolute beginners with a storytelling narrative throughout. Each part follows Ava's space adventure while teaching a core Python concept, includes tables for clarity, code examples, and a hands-on mini-project.

11/11
0 conceptual

Master Outline

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

1 conceptual

Part 1: Welcome to Planet Python — Your First Mission

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

2 conceptual

Part 2: The Planet of Memory Vaults — Variables and Data Types

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

3 conceptual

Part 3: The Crossroads of the Cosmos — Making Decisions

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

4 conceptual

Part 4: The Looping Nebula — Repeating Actions

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

5 conceptual

Part 5: Mission Control Functions — Reusable Commands

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

6 conceptual

Part 6: The Treasure Chest — Lists, Tuples, and Strings

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

7 conceptual

Part 7: The Alien Encyclopedia — Dictionaries

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

8 conceptual

Part 8: The Captain's Log — Reading and Writing Files

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

9 conceptual

Part 9: Space Shields Up — Handling Errors

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

10 conceptual

Part 10: The Grand Mission — Capstone Project

The Adventures of Ava the Astronaut: Learning Python One Planet at a Time

Java Evolution: From 8 to 25

A comprehensive, release-by-release tour of every major feature added to Java from version 8 through version 25. Each part covers one release — explaining the concepts behind each feature, showing practical before-and-after code examples, and placing the change in the context of the language's evolution. Whether you are upgrading a legacy codebase or simply want to stay current with modern Java, this series gives you the depth and clarity to understand not just what changed, but why.

11/11
1 conceptual

Java 8: The Functional Revolution

Java 8 was the most transformative release in the language's history — introducing lambda expressions, the Stream API, Optional, default interface methods, and a brand-new Date/Time API. This part covers every major feature with deep conceptual explanations and practical examples.

2 conceptual

Java 9: Modules, JShell, and a Smarter Platform

Java 9 introduced Project Jigsaw — the Java Platform Module System — fundamentally restructuring the JDK itself. It also brought JShell (the first REPL for Java), convenient collection factory methods, a new HTTP client, and a raft of API improvements.

3 conceptual

Java 10: Local-Variable Type Inference with var

Java 10 was a focused release, best known for introducing the var keyword — local-variable type inference. While small in scope, var meaningfully reduced verbosity in everyday Java code. This part covers var in depth, along with the other improvements in Java 10.

4 conceptual

Java 11: The New LTS Baseline

Java 11 became the new long-term support baseline, replacing Java 8 for most enterprises. It standardised the HTTP Client, added powerful String and Files utility methods, allowed var in lambda parameters, and introduced the experimental Z Garbage Collector.

5 conceptual

Java 12 & 13: Switch Expressions and Text Blocks Begin

Java 12 and 13 were transitional releases that introduced two features in preview that would become cornerstones of modern Java — switch expressions and text blocks (multiline strings). Java 12 also added the Teeing collector and Compact Number Formatting.

6 conceptual

Java 14 & 15: Records, Text Blocks, and Sealed Classes

Java 14 finalised switch expressions and previewed records and pattern matching for instanceof. Java 15 made text blocks production-ready, introduced sealed classes in preview, and delivered the long-awaited helpful NullPointerExceptions. Together they laid the groundwork for modern Java's data-oriented programming model.

7 conceptual

Java 16 & 17: The Modern LTS Baseline

Java 16 finalised records and pattern matching for instanceof. Java 17 became the new LTS release, finalising sealed classes and introducing pattern matching for switch in preview. Together they define the modern Java baseline that most teams target today.

8 conceptual

Java 18, 19 & 20: Virtual Threads on the Horizon

Java 18, 19, and 20 were bridge releases between the Java 17 and Java 21 LTS versions. They introduced UTF-8 as the default charset, a simple built-in web server, and — most significantly — previewed virtual threads, structured concurrency, and record patterns that would be finalised in Java 21.

9 conceptual

Java 21: The Concurrency Revolution — LTS

Java 21 is a landmark LTS release. It finalised virtual threads (Project Loom), sequenced collections, record patterns, and pattern matching for switch — while previewing string templates and unnamed classes. It is the most feature-rich LTS release since Java 8.

10 conceptual

Java 22 & 23: Unnamed Variables and Foreign Memory

Java 22 finalised the Foreign Function and Memory API and introduced unnamed variables and patterns. Java 23 continued refining previews including primitive types in patterns and module import declarations. Together they represent the platform's ongoing maturation.

11 conceptual

Java 24 & 25: The Next LTS — Simpler, Faster, Safer

Java 24 delivered ahead-of-time class loading and stream gatherers. Java 25 is the next LTS release, finalising scoped values, compact source files, flexible constructor bodies, module import declarations, and primitive type patterns — while introducing compact object headers for significant memory savings.

The Azure Ascent: A Backend Engineer's Journey to Cloud Mastery

A comprehensive 6-part tutorial series designed for experienced backend engineers transitioning to Azure. Follow Marcus, a seasoned Java engineer, as he builds CloudVault—a fintech platform—from the ground up on Azure. Each part solves a real production challenge: deploying scalable APIs, designing multi-tier data architectures, securing networks, integrating AI capabilities, and automating everything with DevOps. Combining deep conceptual explanations with 40+ production-ready Java code examples, this series takes you from Azure fundamentals through enterprise-grade deployment practices.

7/7
0 conceptual

The Azure Ascent: A Backend Engineer's Journey to Cloud Mastery

Welcome to the Azure Ascent—a comprehensive 6-part tutorial series for experienced backend engineers. Follow Marcus as he builds CloudVault, a fintech platform, from the ground up on Azure. This series combines deep conceptual explanations with 40+ production-ready Java code examples, taking you from fundamentals through enterprise-grade deployment.

1 conceptual

Part 1: The Summit Awaits — Azure Fundamentals & Core Concepts

Marcus begins his Azure journey by understanding the platform's organizational model. We explore subscriptions, resource groups, regions, and the Azure Resource Manager. Learn how Azure's architecture differs from AWS and why these foundational concepts matter for your designs.

2 conceptual

Part 2: Building the Engine — Azure Compute Services

CloudVault needs to deploy its microservices. Marcus explores four compute options—Virtual Machines, App Service, Azure Functions, and Containers. We build a production-ready Java Spring Boot API and learn when to use each service for different workloads.

3 conceptual

Part 3: Storing the Treasures — Azure Data & Storage Services

CloudVault's data is growing fast. Marcus explores Azure's data services—SQL Database for relational data, Cosmos DB for globally distributed workloads, and Blob Storage for files. We design a multi-tier data architecture and implement production-ready Java code for each service.

4 conceptual

Part 4: Connecting the Dots — Azure Networking & Security

Marcus designs a secure, multi-tier architecture for CloudVault. We explore Virtual Networks, Network Security Groups, Azure Key Vault for secrets management, and identity/access control with Azure AD. Learn how to design networks that protect your data while enabling communication between services.

5 conceptual

Part 5: The Intelligent Layer — Azure AI & Machine Learning

CloudVault wants to add intelligent features to their platform. Marcus explores Azure AI Services (Vision, Language, Speech), Azure OpenAI for LLMs, and Azure Machine Learning. We integrate AI capabilities into a Java backend application and understand when to use pre-built APIs vs. custom models.

6 conceptual

Part 6: Automating the Climb — Azure DevOps & Deployment

Marcus faces his final challenge—automate everything. We build a complete CI/CD pipeline using GitHub Actions, implement Infrastructure as Code with Bicep, set up monitoring with Application Insights, and establish production-ready deployment practices including blue-green and canary deployments.


// career

Experience

2023 — present Tesco

SDE 3 — Technical Architect

JavaScalaSpring WebFluxKafkaCouchbaseKubernetesK3sPrometheusOpenTelemetryApigeeAkamai
  • Leading architecture for the video platform supporting all express stores (~1,500 stores, ~45,000 cameras), with a roadmap to large-format stores — runs on edge K3s clusters with selective cloud export
  • Driving the Total Loss Recovery (TLR) programme — designing systems to link video evidence with loss events for recovery and prosecution workflows
  • Proposed and leading the camera analytics initiative — exploring ONVIF Profile T/S to extract IVA analytics from Hanwha cameras at the edge; designed multi-strategy collection architecture (PullPoint subscription, RTSP metadata parsing, FTP fallback) under 1 Mbps bandwidth constraints
  • Owned architecture from inception for the Incident Prevention platform; led design and implementation of core microservices across a distributed engineering team
  • Designed and built real-time alerting using SSE with Spring WebFlux and Kafka — delivering instant browser notifications to ~200 security hub operators for ~1.5M ANPR events/day and ~3,000 daily fraudulent till transactions
  • Architected the incident reporting system for capturing, correlating, and utilising store incidents (theft, assault) with linked evidence chains for case building and prosecution support
  • Mentor engineers, conduct architecture reviews, and run technical skill-enhancement programmes for the team
2022 — 2023 VMware

Staff Engineer

JavaScalaSparkDynamoDBKafkaAWS
  • Designed and led CloudHealth — a multi-cloud SaaS platform with highly scalable event-driven microservices, integrating a data pipeline processing ~100 TB/day across modules for cloud cost management and optimisation
  • Drove cross-module architectural decisions optimising data flow for low-latency processing and cost efficiency at petabyte scale
2020 — 2022 Bazaarvoice

Staff Software Engineer

JavaScalaFinatraElasticSearchCassandraDynamoDBAWS ECSKinesis
  • Led the Curalate content management platform team — built and enabled a team to take over core high-traffic components from a newly acquired startup; platform handled 3 billion+ API requests during 36-hour Black Friday peak windows
  • Designed and built media processing platform handling ~35,000 images/day, replacing the legacy image processing system with a modern, scalable microservices architecture
  • Managed end-to-end ownership transition, delivery planning, and team mentoring across multiple engineering pods
2019 — 2020 Altimetrik

Senior Software Engineer

Spring BootOracleAngularMongoDBSpring SleuthZipkin
  • Designed and implemented high-performance microservices for asynchronous and synchronous processing of customer data across multiple enterprise systems
  • Implemented distributed tracing in a multithreaded environment using Spring Sleuth and Zipkin for log correlation, performance monitoring, and visual tracing
2016 — 2019 IBM

Advisory Systems Analyst

JavaSpringEnterprise Integration
  • Designed and developed enterprise applications for clients in power utilities and airline hospitality domains
  • Led a small team handling end-to-end delivery from requirements through deployment
2008 — 2016 Earlier Roles

Engineer → Senior Engineer

JavaSpringJPAZK
  • Cognizant (2014–2016): Enterprise applications with Java and Spring
  • Siemens (2011–2014): Core backend modules for industrial products
  • Infosys (2008–2011): Foundational Java and enterprise web technologies

// education

2024 — 2025

Executive PGP in AI & Machine Learning

IIIT Bangalore — Completed coursework and projects; awaiting certificate

2004 — 2008

Bachelor of Technology, Information Technology

Jadavpur University, Kolkata

The Story

Sixteen years of building enterprise systems have taught me one thing — architecture is about trade-offs, not tools. I've shipped petabyte-scale data pipelines at VMware, real-time fraud detection at Tesco, and high-traffic content platforms at Bazaarvoice.

Now I'm at an inflection point. AI isn't replacing architects — it's creating a new class of systems that need architectural thinking more than ever. Production RAG isn't a tutorial problem. Agent orchestration isn't a prompt engineering trick. These are distributed systems problems, and that's where my 16 years become relevant.

"AI: Through an Architect's Lens" is my way of bridging that gap publicly — learning deeply, writing rigorously, and building production-grade systems. Because I believe work speaks louder than credentials.

// titas.config
location "Bangalore, India"
experience 16
primary "Java · Kotlin · Spring"
ai_stack "Python · Haystack · LangGraph"
mobile "Flutter · Dart"
frontend "Svelte · SvelteKit"
education "B.Tech — Jadavpur University"
status "open_to_opportunities"
🧠

Phoenix

Ready when you are

Hey! I'm Phoenix — I know Titas's work, projects, and experience. Ask me anything — from distributed systems to production RAG, or what it's like building at Tesco and VMware.