Case Studies

---

• Low confidence in SW stack leading to need for rigorous testing, validation, and customer-driven improvements.
• Complex and uncertain deployment of first-of-a kind, next-gen robotics in fulfillment centers.
• High-risk launch processes requiring efficiency optimization.
• Scaling challenges in infrastructure, storage, and robotic fleet.
• Inefficient processing times impacting fulfillment speed.
• Knowledge gaps in operations teams on new technology and workflows.

• Conducted 100+ verification & validation tests; integrated customer feedback.
• Scaled the solution from 5 to 55 HMIs, 200 to 1300+ robots, and 40K to 360K sq. ft.
• Deployed advanced robotics in two fulfillment centers within a year.
• Optimized launch plans, cutting critical path duration by 20%.
• Reduced processing time per item from 22 to 10 seconds.
• Provided hands-on training and developed user-friendly materials.

• Faster, more efficient robotic deployment.
• Improved reliability through rigorous testing.
• Minimized launch risks and delays.
• Scalable solutions meeting growing demands.
• Enhanced processing speed and fulfillment efficiency.
• Empowered teams with better tools and training.

• High-risk cost exposure in a $400M+ global CPV project (valve quantity ~35,000 valves) due to high temperature, high pressure and high corrosion conditions in the field.
• Convoluted supply chain among 10 suppliers, across multiple geographies and multiple time zones.
• Unclear priorities for different types of valves leading to exhaustive testing, validation and certification campaigns.
• Cost overruns and inventory inefficiencies.
• Unreliable on-time delivery, and negative cash flow.

• Established 7 satellite teams and one central risk management team, trained over 200 project team members and conducted qualitative and quantitative risk analysis.
• Streamlined global project execution for CPV delivery across major regions (Gulf of Mexico, Angola, Papua New Guinea).
• Optimized supply chain efficiency, reducing cost exposure by 20% and inventory by 13%.
• Implemented a robust business plan ensuring on-time delivery (95%), profitability (12%), and customer satisfaction (>95%).

• Reduced financial and operational risks.
• Increased supply chain reliability and efficiency.
• Minimized costs while maintaining high-quality standards.
• Ensured smooth execution with timely deliveries and optimized cash flow.
• Strengthened customer confidence through effective project management.

• Low volume of peptide was getting manufactured per production shift.
• Unreliable and random quality and chemistry of peptides was generated.
• Lack of automation in the manufacturing process created many bottle-necks in plant operations.
• Unclear throughput, unreliable budget and resource allocation in manufacturing operations.

• Designed, developed and launched a prototype that controlled both pressure and temperature via DAQ and control system in 4 months.
• Optimized the thermodynamics of the process and reduced the drying time of peptides from 18 hours to about 2 hours.
• Increased the throughput of production shift by over 300%.
• Implemented a web-interface, automated the process, and reduced the decision making process duration by over 50%.
• Increased the quality and chemistry of peptides to >95% acceptance criteria.

• Faster decision-making and improved throughput.
• Real-time transparency into product quality and chemistry.
• Minimized risks, increased automation and accuracy in prediction of budget.
• Increase in safety stock to cater surge in demands in different geographies.

• No structured system for collecting and prioritizing requirements from 14 customer teams.
• Inefficient decision-making process (duration = 6+ weeks) compounding delays on downstream processes.
• Lack of visibility for customers, leadership, and internal teams on requirement status.
• Unclear product roadmap and release cadence for satellite software teams.
• Patchy deployment for Flight Control Software (FCS) across production and prototype satellites.
• Technical and schedule risks, along with unresolved trade-offs and blocking issues.
• Gaps in user onboarding, communication, and training.

• Built a Jira and Confluence-based scalable mechanism to collect, evaluate, and prioritize customer requirements.
• Launched a real-time dashboard, reducing decision-making time from 6+ weeks to 1 week.
• Created and implemented a structured product roadmap and release cadence for three satellite software teams– increased the team velocity by 35%.
• Deployed FCS on the first 27 production satellites and provided technical support for two prototypes.
• Led program execution, resolving technical trade-offs, managing risks, and driving change management and reduced technical debt by 15%.
• Designed user onboarding, communication, and training plans for seamless adoption and increased user adoption by 40%.

• Faster decision-making and improved agility.
• Real-time transparency into project status for all stakeholders.
• Streamlined and predictable software release process.
• Reliable and timely FCS deployment for mission success.
• Minimized risks and efficient issue resolution.
• Enhanced user adoption with structured training and onboarding.

• Lack of a structured system to complete Risk Assessment of Radiology Medical Devices.
• Challenging verification and validation of integrating Hardware and Software Radiology Medical Devices.
• Complex Organizational structure with lack of agreement on the commercial readiness of a medical device.
• Misalignment on the acceptance or success criteria for a safe and effective medical device.

• Designed and implemented a scalable framework utilizing Product Development Safety Plans that supported introduction and enhancement of multiple medical devices with similar architecture, thereby saving approximately 100 resource hours over a year.
• Developed a Hazard and Harm Identification (HHI) which led to 15% reduction in overall time spent in harmonization of terminology across cross functional team members. HHI document formed the basis of Risk Assessment for new products. 
• Utilized ISO-14971 to develop a compliant Medical Device Risk Assessment framework.
• Collaborated with various Subject Matter Experts (SMEs) e.g. Systems Engineers, Clinical Practitioners and Quality Engineers in review of Product Safety Risk Assessments and evaluation of technical issues identified during Verification and Validation testing of the products.
• Led Health Risk Assessments to evaluate approximately 25 field issues in 2 years and provided recommendations to the program management team(s). This evaluation provided objective data needed by the Program Management team in identification and resolution of critical issues in the field.

• Faster decision-making and improved agility.
• Structured and consistent approach to Safety Risk Assessment across all Radiology Medical Devices.
• Efficient allocation of resources in completing Risk Assessments for fielded products.

• Energy industry companies spend millions of hours on collaboration with partners including Engineering, Procurement, and Construction (EPC) contractors.
• The collaboration workflows include reviewing project documentation, answering technical questions, searching and tracing information, and auditing changes.
• Companies review thousands of such documents per year. Millions of man-hours of Engineering personnel are wasted due to ineffective collaboration. 
• Lack of traceability and non-standardized audit leading to errors accumulating in the review of project-critical documents.

• Developed a web-based application for collaborating with EPC companies that scaled to 5 capital projects with over 1,000 users.
• Reduced review time for engineers for document review and question-answer workflows by 40%.
• Efficient workflow routing improved administrative efficiency by 20%.

• Reduced collaboration challenges and streamlined operations for capital projects.
• Integrated the solution with existing information management systems of the customer to enable a seamless transition to new workflows.
• Algorithm-based routing of documents aligned closely with the customer's organization hierarchy, delegation policies, and project level roles and responsibilities.
• Effective automation in the workflows enabled auditability across organizations and large scale capital projects.

• Lack of a unified platform to integrate 35+ sensors (Lidars, radars, cameras), AI processors and control systems
• Challenges in converting the traditional automotive plant operations into autonomous vehicle manufacturing
• Lack of geofence area to validate autonomous fleet under a diverse real-world scenarios

• Established strategic partnership with an AI start-up for an autonomous software development and a simulated platform for continuous learning
• Outsourced non-automotive components to Tier-1 suppliers and improved development cycle by 18 months
• Partnered with a small size manufacturer for a flexible manufacturing line using a scalable, cost effective manufacturing workflows and reduced production cost by 40%.
• Conducted safety validation scenarios using Hardware in loop system and reduced the 1st prototype pedigree fleet size by 25%. 
• PP&C solution integrating multiple partners to achieve agile product development and reduced time to market by 24 months. 

• Built an autonomous vehicle software with 100M+ miles of real world and simulated data
• Introduced modular assembly line for autonomous vehicle production
• Delivered the first fleet of 500 vehicles within 5 yrs of development cycle
• Introduced SAE level 4 autonomous car for ride hailing services in the 2 U.S. cities.

• 6+ months launch delay of an Electric motor by Tier1 automotive supplier jeopardizing the start of production (SOP) of Electric Vehicle 

• Reduced the average implementation time for changes from 15 days to just 3 days with accelerated processes for change management 
• Achieved 2 months recovery by organizing onsite workshops to correct BOM issues and resolve supplier issues.
• Satisfied customer intermediate milestones with additional prototype samples. 
• Achieved schedule improvement by managing staggered launch of the long lead components. 

• Reduced material lead time and improved parts availability meeting customer’s Material Required Date milestone
• Protected customer Start of Production milestone with on-time delivery of the Electric Motors. 

 Problems and Issues

Facility reports provide critical operational information in the energy industry. Facilities generate these reports daily, weekly, monthly, and quarterly, creating a massive volume of data. Manual processing of these reports is time-consuming and prone to human error, leading to inaccurate or missing data. Our customers need an efficient solution to process these reports at scale to gain valuable operational insights.

We implemented an AI-enabled document processing solution that achieved 95% accuracy in extracting key actionable insights from facility reports. Our system utilizes optical character recognition (OCR) and machine learning algorithms to process over 2 Million pages of reports in just 5 hours. Through analysis of Non-Productive Time (NPT), we helped reduce operational waste by 10%.

• Enabled data-driven decision making through comparative analysis against industry standards
• Significantly reduced report processing time and costs
• Enhanced data accuracy for identifying health hazards across facilities
• Improved safety metrics through better tracking of near-misses and incidents
• Enabled implementation of data-backed safety policies

Leading Medical Device manufacturer sells over 60 million disposable kits to be used with power injectors in hospitals across the work. These disposable kits are used to inject contrast / saline into patients for diagnostic imaging.

These disposable kits are assembled and packaged using highly automated syringe assembly lines. Despite robust quality control processes manufacturing defects may accidentally get introduced in products released into the field.

Key challenges included:

• Delayed detection of occasional manufacturing defects into the finished product
• Customer dissatisfaction and disruption in patient care due to failed product
• Manual data triage, increasing operational costs
• Lack of predictive insights into process variations

With the use of Generative AI and prompt engineering, we performed the data analysis to identify the root cause in a time efficient and cost effective manner. The approach included:

• Customer complaint data was pre-classified based on various allegations reported by the customers and segmented based on risk to the patients or users.

• Using Generative AI and Prompt engineering, a time series analysis was completed on the unstructured data and a trend was identified indicating seasonality of the defects. With Generative AI, trend identification saw order of magnitude improvement from 24 hours to 20 minutes. This contributed to more appropriate usage of resources within the team.

• Based on the defect identified and its seasonality, Generative AI and prompt engineering completed by Subject Matter Experts, most likely root causes was identified. 
• These results indicated most likely root causes saved approximately $15K in outside material analysis. With the use of Generative in AI identifying preliminary root causes for defect has the potential to save thousands of dollars in exploratory testing for preliminary complaint sample investigations.

• Manufacturing and assembly process parameters were reviewed based on the test results and identified material drying process as the process of interest.
• The material drying process was found to be operating outside of the recommended parameters and was adjusted as needed.
• Routine maintenance procedures were updated to ensure processes are closely monitored thereby minimizing chances of future defect.

Increased Customer Satisfaction:

• Customers impacted by the defect were appreciative of quick turnaround leading to higher Net Promoter Score (NPS)

Field defects caused by quality issues were identified and isolated with least amount of delay leading to quick root cause analysis and process corrections as needed.