Technology
Navigating EV Manufacturing Quality with Karthick Ravichandran
Karthick Ravichandran’s core argument is one the automotive industry is learning under pressure: quality planning ...
With over two decades of experience leading engineering and program management at Rivian, Honda R&D, and Alps Alpine, Venkata Repaka is pioneering the integration of AI into automotive R&D to drive intelligent cost optimization. His Intelligent Value Engineering framework combines traditional methods like DFMEA and DFM/DFA with predictive analytics to reduce lifecycle costs and accelerate innovation. By bridging design, data, and decision-making, Repaka is shaping a new era where program managers act as AI orchestrators—blending human expertise with machine intelligence to create smarter, more efficient mobility solutions.
The global automotive industry faces an unprecedented period of transformation, balancing significant capital demands brought by technological innovation with declining profitability. Automotive suppliers, in particular, find themselves under immense pressure, with earnings margins persistently below pre-pandemic levels.
This scenario represents more than just a temporary downturn; it’s a profound structural shift driven by stagnant global vehicle sales, slower consumer adoption of battery electric vehicles (BEVs), and escalating cost inflation. Additionally, heightened geopolitical trade friction and new tariffs further threaten critical investments in R&D precisely when innovation is most needed.
Navigating these challenges calls for a novel approach to value creation, a challenge embraced by Venkata N Chandra Sekhar Repaka, an automotive engineering and program management expert with over 20 years of experience in vehicle interior design, closures, and complex systems integration. Having led cross-functional teams at prominent organizations such as Rivian, Alps Alpine, and Honda R&D, Repaka specializes in engineering design validation, tool development, and system optimization.
With significant experience in CAD/CAE approaches, DFMEA, and Design Validation Plans & Reports (DVP&R) and Six Sigma Green Belt certification, he is a votary of a pathbreaking paradigm of Intelligent Value Engineering.
Repaka’s innovative strategy goes beyond traditional reactionary cost-reduction strategies by embedding artificial intelligence into the very research and development workflows. He is focusing on embedding AI into the workflows of vehicle design, highlighting platform-based architectures and software-defined vehicles.
By combining machine learning and high-end analytics with the time-honored traditions of the engineering discipline, like DFMEA and Design for Manufacture and Assembly (DFM/DFA) design, he espouses a predictive, data-based method of automotive product development.
By way of in-depth coverage of AI incorporation in core engineering workflows, Repaka delineates the way resource limitations as well as technological intricacies can be managed by automotive groups effectively. It is not just a guide on how current economic tailwinds could be overcome, but also the harbinger that could bring better efficiency and competition, gaining an advantage.
The modern automotive sector operates under a state of relentless financial pressure, where the need to invest billions in next-generation technologies collides with declining profitability. An established original equipment manufacturer (OEM) might need to commit over $70 billion just to establish a defensible position in these new trends. This environment demands a fundamental shift in how value is created and costs are managed, moving the focus from downstream reduction to upstream prevention.
Repaka’s perspective is shaped by years on the front lines of this challenge. He states, “My career in automotive engineering and program management has granted me deep insight into the relentless cost pressures and complexities of R&D. Customers and senior management expect consistent year-over-year savings, and early-stage R&D decisions have an outsized impact on downstream costs.”
This latter point is a well-documented principle; research confirms that decisions made during the initial design phase can lock in up to 80% of a product’s total lifecycle cost. Changes made late in the development cycle can cost as much as thirteen times more than those made upfront, a result of the cascading rework required across tooling, testing, and supplier networks.
This is where AI emerges as a strategic enabler, capable of bridging the gap between the high-level business objective of cost reduction and the complex, technical trade-offs of engineering. By embedding AI-powered analytics into the earliest stages of development, teams can gain proactive insights into the cost implications of their choices.
“My technical expertise allows me to understand engineering trade-offs, while my program management experience ensures that we are meeting business objectives,” Repaka explains. “In this context, AI becomes the scalable bridge between innovation and execution, transforming raw data into actionable strategies that drive efficiency at every stage.”
To effectively leverage AI for cost optimization, organizations must first identify and harness the most valuable data sources. In automotive product development, this means breaking down traditional data silos to create a unified view that connects design intent with real-world cost drivers. This holistic data strategy is the bedrock of any successful intelligent value engineering framework.
In Repaka’s view, the strongest insights are the ones that are derived from the combination of multiple, non-correlated datasets. “The best data sets for AI-based cost reduction in automotive product development that I know are the ones that connect design, manufacture, supplier, and performance facets,” he says.
“The secret is linking technical data with cost drivers at the outset of the R&D process,” he says. It entails the establishment of a comprehensive data ecosystem encompassing design and engineering data, like CAD files and Bills of Materials (BOMs); cost and procurement data, such as supplier quotes and raw material price trajectories; and test and manufacturing data.
AI-based tools can today automate much of the integration work, and certain tools can instantly decrease document handling time by as much as 90% through automated reading of the specifications out of the CAD document.
It is integrated data that fuels predictive cost models that can accurately forecast new part costs in real time, a giant leap beyond the methods of the past. Repaka points out that certain high-impact variables are key in this ecosystem.
“I use should-cost models, which offer benchmarked part and part costs by geography or supplier, and supplier quotes and minimum order quantities,” he observes. “Raw material price trends in metal, polymer, and coating are very important as well.”
By learning machine learning algorithms on these globally connected data points, AI computer programs can excavate the root relationships between design elements and the corresponding cost, allowing the engineers instant results on the monetary effects of the decision.
Implementation of AI is not a replacement, but a supplementation, that is, supplementing time-proven methods with data-based insights. An efficacious paradigm of smart value engineering is not replacement, but completion, that is, completing the systematic direction of methods such as DFMEA with the velocity as well as the predictive capability of AI.
Those root methods offer a fundamental, disciplined way into risk and manufacturability. As Repaka says, “Time-worn methods like DFMEA and DFM/DFA offer systematic, empirical advice on design risk, manufacturability, and assembly efficiency. They are vetted under fire and indispensable for cross-functional agreement and regulatory ability.”
They are the agreed-upon language that engineering teams use when determining possible failure points and making certain that a product can be reliably and efficiently built. They are frequently, though, limited by the team’s knowledge and often have to be done manually, something that AI is just the right size to plug into.
By integrating these activities with AI, companies can take them from being stodgy, paper-based exercises to being dynamic, computer-based intelligence. For example, AI-driven software can scan enormous historical databases of previous failure modes and vigilantly detect new risks in a new design, agreeing with human masters over 90% of the time, but at a very high speed.
“AI never replaces traditional methodologies—it complements them by bringing data-richness and the capability to learn from past as well as real-time data constantly,” says Repaka. “I do use DFMEA and DFM as a base and then complement them with AI-based insights to offer precision, velocity, and foresight.”
In spite of the revolutionary possibilities, the incorporation of AI into existing auto R&D units spells high challenges. Roadmap progress towards intelligent value engineering is frequently stymied by technical as well as cultural roadblocks, ranging from disjointed data environments to skepticism. Overcoming these challenges demands an intentional tack that revolves around establishing a robust data foundation as well as the creation of credibility through transparency.
One of the leading technical barriers is the condition of the data that is organized. In large enterprises, information that is critical is found fragmented across legacy systems and trapped inside departmental data silos, which complicates the process of training robust AI models.
Repaka points out the key issue as data silos and incompatibility, saying, “A key challenge is data silos and incompatibility, with key design data from design, cost, test, and supplier teams co-existing in distinct systems. We solve this by agreeing on and aligning part metadata upfront, allowing efficient AI training and cross-functional analysis.”
Just as critical is the defeat of cultural resistance among engineers who, correctly, are fearful of “black box” AI systems whose reasoning logic is obscure. In a safety-critical industry, verifiability and accountability always trump.
“One other challenge is that engineering skepticism towards AI, as the team will hesitate to put faith in ‘black-box’ recommendations,” says Repaka. “I work on explainable AI—utilizing tools such as the SHAP plots and visual logic trees, making decisions clear and believable.”
Explainable AI (XAI) is a key discipline that offers techniques to interpret and comprehend the rationale underlying the output of an AI. By making the logic behind the model clear, the XAI enables the engineers needed trust such that they can embrace these tools, making the AI go from being an unverified oracle to a verifiable and trustworthy cooperator.
Living-hinge fasteners are widely used in automotive and industrial applications for their low weight and easy assembly. However, recurring issues—such as lid separation and hinge breakage during transportation and molding—have increased customer rejections and reduced component reliability. This paper presents a novel pin-and-socket fastener that replaces the conventional living hinge and retention tab to avoid self-locking, addressing key durability and manufacturability challenges.
Traditional living-hinge fasteners rely on thin polymer connections between the base and lid. These features are prone to fatigue and deformation during handling and transportation. Across multiple production batches, lids were observed separating from the base or cracking during shipment, resulting in high rejection rates and additional rework costs. A redesign was required to increase hinge robustness, eliminate cracking, and maintain manufacturability without raising part cost.
Field feedback indicated that hinge failures were primarily caused by:
These issues led to visible damage and lid separation, resulting in product nonconformance.
The new design introduces a pin-and-socket locking mechanism that structurally links the lid and base, eliminating dependence on a flexible hinge.
Key elements include:
Testing and production validation demonstrated that the redesigned fastener:
The proposed pin-and-socket fastener offers a robust, manufacturable, and cost-effective alternative to conventional living-hinge mechanisms. By mitigating failure modes related to cracking and premature lid closure, the solution improves both product reliability and process efficiency. The design has been successfully implemented in production, contributing to higher customer satisfaction and reduced warranty concerns.
Transformations cannot be powered by technology alone. Intelligent value engineering success is predicated on human collaboration. Inseminating this data-centric mindset throughout an organization demands dismantling the customary silos between the like-engineering, sourcing, manufacturing, and quality departments. Only when all the stakeholders have one goal in common and one common perspective on the data is there true alignment.
This collaborative culture is the human infrastructure that underpins the technology. Repaka stresses that co-ownership is the recipe for success.
“Value engineering works when all parties—engineering, sourcing, production, quality, and suppliers—can look into the solution and say, ‘That’s me.’ My work is aligning that,” he says.
Isolation is the condition under which teams work with misaligned priorities and a lack of information, causing inefficiency and expensive rework. But when teams are functioning as efficient, cross-functional teams, communication is enhanced, problems are better solved, and decisions happen faster. It is this intentional creation of a collaborative culture that enables the insights that AI produces to be translated into actual business results.
As the automotive industry continues its profound transformation, the role of AI in R&D will only intensify, reshaping not just the tools engineers use but the very structure of leadership and the skills required to succeed. The automotive AI market is projected to grow exponentially, reaching nearly $50 billion by 2034 as it becomes the central nervous system for everything from design to manufacturing. This technological sea change will demand a new generation of leaders and engineers who can effectively orchestrate a hybrid ecosystem of human talent and artificial intelligence.
The traditional role of the program manager is set to evolve significantly. As AI automates routine tasks and provides predictive insights, leaders will shift their focus from managing tasks to designing and overseeing complex, interconnected systems.
“Looking ahead, Program Managers and VAVE leaders will become AI orchestrators, not just task managers,” Repaka predicts. This emerging role of the “AI Orchestrator” involves managing a network of specialized AI models, data pipelines, and human experts to function as a single, cohesive unit.
This new paradigm further requires a new engineering workforce with a new skill base. Although the core domain expertise is still key, the same should be complemented by data science and software expertise.
“The highest ROI will be derived from the development of hybrid engineering teams that know data and design intent analysts who know design intent,” says Repaka. Most valuable professionals will be engineers who can work at the intersection points between areas—engineers who know how to pose the correct questions to an AI, critically assess the output, and imaginatively use the output as insights to develop innovative solutions for hard-to-solve problems.
This shift is one towards non-siloed departments and towards agile, mission-based teams that can handle the complex issues that the next era of mobility will have.
In an industry characterized by low-cost pressures and accelerating technological disruption, the centuries-old approaches to automotive R&D no longer apply. Competitiveness in the next decade will be as much about the intelligence and efficiency of the process by which things are made as the things that are made.
The paradigm of Intelligent Value Engineering, one that blends intensive subject matter expertise with the computational capability of artificial intelligence, presents a clear and tangible way forward. By shifting foundation-level processes such as DFMEA and cost studies into flexible, forecasting-based systems, this method integrates cost avoidance and risk defense into the very marrow of the design process.
It transforms the engineers’ and program managers’ tasks into strategic orchestrators of a hybrid human-AI workforce. By embracing this paradigm shift, the automakers could not just survive the gigantic challenge of the current times but also open a new dimension of innovation, efficiency, and long-term value creation towards the future of mobility.
