Skip to content

Applus-poc

Project name: Applus-poc🔗

Deep Kernel Labs (Web page)

Team:🔗
Applus
Client:🔗
Applus
Role:🔗
Software Engineer | Data Engineer

April, 2026 - June, 2026

Ourense, Spain

Recommendation: See on Linkedin

This project involved developing a proof-of-concept (POC) web platform for the accredited certification authority, Applus. The platform helps to standardize inspection processes, create new inspection workflows from regulatory standards using AI assistance, visualize the regulatory sources referenced in each inspection process, interact with an AI-powered chatbot to answer questions about inspection procedures, and migrate various Applus inspection process documents to the Fieldeas format. Contributed to the platform's architecture and implementation by designing modules and class structures, leveraging AI to generate implementation code, and validating the generated code through comprehensive edge-case testing to ensure correctness and reliability.

Hard / Technical achievements🔗

  • Contributed to the design of the platform's architecture, collaborating on the overall system structure and technical decisions.
  • Designed and implemented, using Python, React and AI assisted development; a feature that enables users to visualize the regulatory sources referenced at each step of an inspection process, providing direct traceability between inspection procedures and the corresponding regulatory requirements.
  • Extended, using Python, docling and AI assisted development; a library that were used for extract document sections into one that extract and convert sections of regulatory documents into HTML, enabling large language models (LLMs) to process the information more efficiently and accurately.
  • Leveraged AI-assisted development workflows for brainstorming, system design, implementation, testing, debugging, and bug fixing, improving development efficiency while reducing model hallucinations through effective prompting and validation techniques.
  • Reviewed, tested, and validated AI-generated Python code before integration, ensuring correctness and preventing design and implementation errors resulting from model hallucinations.
  • Used AI to rapidly understand large and unfamiliar codebases, significantly reducing onboarding time and accelerating development on the project.

Soft achievements🔗

  • Attended to daily meetings with the team to coordinate tasks and find potentials problems in the development of the solution.
  • Used Jira to organize and track the assignee tasks.
  • Followed the scrum methodology guaranteeing an agile development process.
  • Exposed in the daily meetings the progress of my tasks, helping to others to understand my progress and the implementation of my tasks.
  • Participated in pair programming sessions with teammates to design the system, define implementation tasks, and help distribute work across the team.

Others team members:

Stack:
Python, Django Rest Framework, JSON, SQL, PostgreSQL, pgAdmin, Claude Code (Opus 4.7), Git, Github, React, Docker, docker-compose, docling (Python library), Api Rest.
keywords:
Agentic programming, AI Engineering, containers.