ESDS Builds an AI-human Bridge for its Software Development …
Analytics India Magazine (Ankush Das)
Artificial Intelligence (AI) is steadily becoming embedded across the software development lifecycle, not as a flashy add-on but as a structured part of engineering operations.
In an exclusive interaction with AIM, Rushikesh Jadhav, chief technology officer of ESDS Software Solution, described how AI now influences testing, deployment, resource planning, and infrastructure behaviour in measurable, if still evolving, ways.
A key part of ESDS’s transformation is its decision to use AI beyond code helpers. This includes automated vulnerability assessment, predictive analytics for resource allocation and decision support, and support for its own cloud platform.
“The move has smoothed workflows, minimised manual interventions, and sped up feature rollouts, although specific metrics such as defect rates are showing qualitative improvements with CMMI Level 5 processes,” Jadhav said.
Where AI Fits, Where Humans Lead
The company sees the strongest returns in testing and deployment, confirms Jadhav. Regression tests now run automatically, and CI or CD (continuous integration and continuous delivery/deployment) systems adjust using AI-assisted tuning. This has reduced repeated manual effort and allowed engineering teams to stabilise release patterns.
“In the SDLC (software development life cycle), AI excels in testing and deployment phases by automating regression tests and optimisations of CI/CD to meet efficiency,” he said.
But ESDS also maintains clear limits. Requirement gathering and architectural design remain human-led because they demand contextual understanding that AI cannot replicate today. The company’s CMMI Level 5 processes reinforce this boundary, ensuring that automation augments engineers rather than overrides them. Even in operations, where AI flags anomalies in logs, engineers remain the final decision-makers.
This balanced approach reflects a broader philosophy where AI is used where efficiency matters, and human judgement comes in where reliability and interpretation are essential.
Treating AI-Suggested Code With the Same Rigour
As AI becomes more involved in code generation, ESDS has reinforced its validation pipeline. Every suggestion, whether human-written or AI-generated, must pass through structured checks.
“For mission-critical applications, we validate AI-generated code by performing multiple layers of review: automated static analysis, peer audits of code, and integration testing in simulated environments,” Jadhav said.
For data centre software, these reviews extend into stress testing to ensure compliance with zero-downtime expectations. AI tools assist with security scanning, but they do not replace human judgment. Every change still requires manual sign-off under ISO 27001 protocols, reinforcing the company’s belief that automation should strengthen, not dilute the engineering discipline.
Data: The Backbone of AI Development
ESDS relies on proprietary operational datasets, including data centre logs, customer usage patterns and a centralised knowledge repository of previous deployments. These datasets power AI systems built into its solutions such as eNlight 360°. Governance remains a central pillar.
“It ensures that high-quality data is maintained, including automatic cleansing and governance frameworks,” Jadhav noted.
Security is tightly enforced through encryption, anonymisation and compliance with general data protection regulation and payment card industry data security standard. This allows the company to train models on real behaviour without risking sensitive information, a critical requirement for enterprise clients that depend on ESDS infrastructure.
Tooling and Culture – The Real Obstacles
Contrary to common industry narratives, the company does not see talent shortages as the primary barrier. Instead, Jadhav pointed to deeper structural issues.
“From where we stand, the biggest challenge is tooling maturity and organisational change,” he said.
Many enterprise environments rely on legacy systems that do not expose data or control signals required for modern AI models. Teams must also adjust to new operating rhythms in which automated tools make frequent, low-level decisions. Building trust in these systems, retraining teams and updating internal processes creates organisational friction that technology cannot alone solve.
One of ESDS’s most effective AI deployments is in its patented eNlight Cloud platform. Predictive models analyse demand patterns to forecast spikes and scale resources automatically before a surge occurs.
“This has reduced unplanned scale-up, which directly equates to quicker onboarding of customers for clients at peak loads,” Jadhav described.
This prevents bottlenecks, reduces latency risks and lowers overall infrastructure costs for customers. It is also a practical demonstration of how AI can improve system behaviour without increasing risk.
A Development Culture Evolving With AI
ESDS’s experience shows that AI in software development is not about replacing engineers but refining systems.
Automation handles routine tasks, predictions strengthen infrastructure resilience, and engineers focus on architectural, ethical and contextual decisions.
With its mix of telemetry-driven models, governance frameworks and human oversight, the company illustrates what a mature AI-enabled development lifecycle can look like.
The post ESDS Builds an AI-human Bridge for its Software Development Lifecycle appeared first on Analytics India Magazine.
Generated by RSStT. The copyright belongs to the original author.