Recursive Zero RECURSIVE ZERO

My Journey Python and AI Intern at Recursive Zero

A reflection on my internship experience at Recursive Zero, working on AI-powered fabric applications, computer vision systems, backend engineering, deployment pipelines, and production-ready machine learning solutions.

Poorna Chandra
#internship #ai #machine-learning #python

My Journey Python and AI Intern at Recursive Zero

From January 2025 to February 2026, I had the opportunity to work as a Software Development Engineer (SDE) Intern at Recursive Zero, an early-stage startup focused on applying AI to fabric analysis and transformation.

One of the most valuable aspects of this internship was the opportunity to work directly with the CTO, Keshav Mohta. Coming into the internship, there was a lot I didn’t know. Mr. Keshav Mohta was incredibly patient throughout the process, always willing to explain context, answer questions, and help me get up to speed. That support made it much easier to contribute effectively and take ownership of increasingly complex work.

Building AI-Powered Fabric Applications

A major focus of my work involved developing systems for fabric analysis and transformation. These projects were not limited to research experiments—they evolved into live applications that were actively deployed and used.

During the internship, I worked on end-to-end computer vision pipelines that analyzed fabric images and performed image transformation workflows. This included data processing, model integration, API development, deployment, and ongoing improvements based on feedback.

The applications were deployed and maintained at:

One of the technical challenges I worked on involved fine-tuning YOLO-based object detection models to identify multiple fabric regions within group images. This required understanding both the machine learning workflow and the practical considerations involved in preparing data and evaluating model performance.

I also implemented multimodal search capabilities using LanceDB and SigLIP embeddings. This allowed users to retrieve relevant results using either images or text, enabling more flexible search experiences. Working with vector databases and embedding models gave me hands-on exposure to modern AI retrieval systems and similarity search techniques.

Developing Production APIs

Beyond machine learning, a large portion of my work involved backend engineering.

I developed FastAPI services for:

  • Machine learning inference
  • OCR processing
  • Image transformation workflows
  • Search operations
  • Data validation
  • Rate limiting

Building Reliable Data Pipelines

Several features required extracting information from images and storing it in structured formats. To support this, I developed OCR-based extraction pipelines that processed image data, validated results, and persisted information into databases.

MongoDB was used for metadata storage, and I worked on schema design, indexing strategies, and efficient data retrieval patterns. I also integrated AWS S3 for image storage, allowing applications to manage and serve image assets reliably.

These experiences helped me understand how machine learning systems interact with storage, databases, and application infrastructure in real-world deployments.

Engineering for Deployment

Considerable effort went into deployment, packaging, dependency management, and optimization. I worked with Python, Poetry, and Streamlit to build production-ready applications while following modular architecture and clean code practices.

Some of the areas I contributed to included:

  • Designing reusable Streamlit project templates
  • Optimizing dependency management and build configurations
  • Reducing deployment artifact sizes
  • Packaging Python modules using pyproject.toml standards
  • Maintaining reproducible development environments

Open Source Contributions

The internship also encouraged collaboration beyond internal projects. I contributed to open-source initiatives through pull requests, code reviews, and tooling improvements. This helped me become more comfortable working within larger codebases and collaborating in structured development workflows.

Key Takeaways

This internship taught me much more than individual technologies. It showed me how machine learning, backend engineering, cloud infrastructure, databases, deployment, and product requirements come together to create systems that deliver real value.

Over the course of the internship, I gained hands-on experience with:

  • Python
  • FastAPI
  • Streamlit
  • LanceDB
  • SigLIP
  • YOLO
  • MongoDB
  • AWS S3
  • OCR systems
  • Cloud deployment workflows

Most importantly, I learned how to move beyond building isolated features and start thinking about complete systems—from development to deployment and maintenance.

Conclusion

My internship at Recursive Zero was an incredibly impactful experience. It provided me with a clear window into the corporate world and taught me how development teams collaborate to build amazing products.

I am immensely grateful to Keshav Mohta and the entire team for their guidance and support throughout this journey. I learned a great deal and look forward to applying these skills in future endeavors.

Poorna Chandra