In late 2025, several teams of Fincons professionals took part in the challenge of designing Agentic AI solutions in the Fincons Agentic AI Hackathon, managed by the Fincons Technology Innovation Hub. Each team, composed by colleagues from our US and India offices, demonstrated exceptional teamwork, skills and creativity in developing advanced artificial intelligence solutions. We interviewed the three teams that chose to focus on the green coding track as this practice is becoming increasingly critical to development as more businesses demand sustainable solutions of their technology partners.
The EcoDev team is composed of members: Riya Jain, Ritik Nakra, Vikas Lavaniya and Sumant Reddy, each bringing their own unique view to the development of a green coding solution for the Fincons’ hackathon.
We asked them to tell us a bit about what they developed and their thoughts green coding. The team produced the CodeGreen.AI solution.
When asked why they selected the green coding track they shared the awareness that there is a very little visibility or standardised green KPIs in everyday development workflows and, as Riya points out: “That gap motivated us to choose the green coding track and build a solution that make sustainability measurable and actionable directly within the coding process.”
When selecting KPIs the team started with a simple, yet effective question Ritik says: “What actually consumes energy in a software system?”. The result is a solution that aims for a reduction in energy consumption of up to 30%, along with a reduction in cloud costs and CO₂ emissions. “Our goal with the CodeGreen.ai is not just to detect the inefficient code, but to help developers understand why something is inefficient and how they can improve it. So thinking of it like having a mentor sitting next to you during your coding,” explains Vikas.
Their solution, CodeGreen.AI is thus a kind of "fitness tracker for coding sustainability", integrating another layer into quality assurance (QA) with the detection of energy inefficiencies.
Specifically, the solution is based on a sequential workflow composed of four main nodes as Sumant explains: “The first node is the input validation that checks whether a submitted code or pull request link is valid or complete. This ensures that the analysis is reliable from the start. The second node is the issue detection and suggestion node. This is where AI analysis is done. The third node measures performance metrics quantitatively, evaluating the code against key performance indicators (KPIs) defined for sustainability such as CPU Efficiency. Finally the fourth note is the report generation node that creates a report for any type of stakeholder.”
Watch the whole interview below to find out more about CodeGreen.AI and the ideas behind it.