What happens when a software engineer with more than 12 years of backend engineering experience starts building applications using AI instead of manually writing code?
In this episode of CoderLegion Developer Spotlight, we sit down with Sudhanshu Ratna Thakur, a software engineer and technical lead from India, to explore his journey from biotechnology to software engineering, his experience with large-scale backend systems, and his growing involvement in AI-powered application development.
From banking systems and Kafka to AI agents and the Model Context Protocol (MCP), this conversation explores how software development is changing and why reliability, security, and human supervision still matter.
🎥 Watch the Full Interview
Guest: Sudhanshu Ratna Thakur
Host: Tom Smith
Series: CoderLegion Developer Spotlight
Watch the full interview on YouTube
👨💻 From Biotechnology to Software Engineering
Sudhanshu's journey into software development didn't begin with a traditional computer science background.
He started with biotechnology, where he became interested in bioinformatics and programming, particularly using C for computational work related to drug discovery and protein analysis.
That early curiosity eventually led him to learn Java and other technologies, explore backend and frontend development, and begin his professional journey in the IT industry.
Today, he has more than 12 years of experience working with backend engineering and distributed systems.
His journey is an interesting reminder that a career in technology can begin from unexpected places.
⚙️ Building Reliable Backend Systems
A major part of our conversation focuses on the challenges of developing reliable backend systems, particularly in the banking industry.
Sudhanshu shares his experience with:
- Java-based backend development.
- Apache Kafka and event-driven systems.
- Transactional outbox patterns.
- Banking transaction processing.
- Request IDs versus event IDs.
- Database reliability and data consistency.
- Banking modernization using cloud technologies.
One important topic is how banking systems need to handle transactions reliably. Missing a transaction or processing something incorrectly can have serious consequences.
He also discusses why proper testing, transaction handling, and system modernization are essential when working with financial applications.
🚀 Building Applications With AI Prompts
One of the most interesting parts of this interview is Sudhanshu's recent experience building a US-based healthcare application using AI.
He explains how he integrated project documentation and architecture information from Confluence into his development workflow.
Using AI prompts, he generated different parts of the application, including:
- Application architecture and design.
- Project folder structure.
- Configuration files.
- Backend APIs.
- Database and AWS integration.
- Testing workflows.
He describes completing this project without manually writing individual lines of code, while still reviewing the generated code and checking whether everything worked correctly.
He also discusses using Postman collections to automate testing scenarios.
This experience raises an interesting question: How much of traditional software development will change as AI becomes more capable?
🤖 AI Agents, MCP, and the Importance of Human Supervision
We also explore Sudhanshu's thoughts on AI agents and their growing role in software applications.
He discusses an MCP-based project involving multiple data sources, including:
- MongoDB.
- Snowflake.
- Excel files and documents.
The conversation touches on data retrieval, correct response mapping, integration testing, and caching.
But Sudhanshu also raises an important concern: AI agents can make mistakes, especially when they have excessive access to databases and other systems.
He believes that AI agents need appropriate permissions, proper controls, and human supervision.
Even as AI becomes more capable, responsibility for its actions cannot simply be ignored.
📊 Other Important Topics From the Interview
Here are some additional subjects covered in our conversation:
- Why transactional outbox patterns matter.
- Request IDs versus event IDs.
- Banking modernization and migration challenges.
- Load testing with JMeter and Gatling.
- Concurrent users and application performance.
- Why average response times can hide performance problems.
- Preparing applications for production.
- QA testing and production readiness.
- The future of AI agents and software engineering.
- Why human accountability remains important.
💡 Advice for Developers Starting Their Careers
At the end of the interview, Sudhanshu shares a simple but valuable message for people beginning their software development careers.
His advice is to stay curious, keep learning every day, and start with small things instead of trying to understand everything at once.
Technology changes quickly, especially with AI becoming part of everyday development workflows. Developers who remain curious and continue learning will be better prepared for what comes next.
👤 Meet Our Guest: Sudhanshu Ratna Thakur
Role: Software Engineer & Technical Lead
Location: India
Experience: 12+ years in backend engineering and distributed systems
CoderLegion Profile:
https://coderlegion.com/user/Sudhanshu.thakur
🎙️ Interview Host: Tom Smith
This episode is hosted by Tom Smith, who leads the conversation about Sudhanshu's professional journey, engineering experiences, and perspectives on AI-powered software development.
CoderLegion Profile:
https://coderlegion.com/user/Tom Smith
🌐 About CoderLegion Developer Spotlight
CoderLegion Developer Spotlight is an interview series featuring software engineers, developers, technical leaders, founders, and independent builders.
Our goal is to give developers a platform to share their real-world experiences, technical knowledge, career journeys, projects, lessons learned, and perspectives on the future of technology.
We believe every developer has a story worth sharing.
Want to Be Featured in a Future Developer Spotlight?
Are you building something interesting? Have you solved a challenging engineering problem, launched a product, explored a new technology, or learned something valuable during your development journey?
We'd love to hear your story.
Connect with the CoderLegion community and explore opportunities to participate in future Developer Spotlight interviews.
🌐 Website: https://coderlegion.com
What do you think?
Can AI eventually build complete production-ready applications without developers manually writing code, or will human supervision always remain an essential part of software engineering?
Share your thoughts in the comments!
DeveloperSpotlight #CoderLegion #SoftwareEngineering #AIAgents #BackendDevelopment #Java #Kafka #MCP #DistributedSystems #ArtificialIntelligence
