Zoolatech

Client Challenge:
Modernize legacy systems to boost scalability and efficiency, avoid downtime, and speed time-to-market—without disrupting mission-critical MES/QMS products. 

Solution:

• Embedded a senior engineering team, removing customer–vendor barriers and enabling proactive idea sharing.

• Built new microservices-based components; learned and extended complex legacy systems quickly.

• Optimized DB/Spring usage, established QA automation, and tightened delivery processes.

Results:
Faster delivery cadence and higher product quality; seamless collaboration enabled quick feature releases and positioned the client for sustained growth in highly regulated industries

PANDORA IS A FAMOUS EUROPEAN JEWELRY MANUFACTURER AND RETAILER SELLING ITS PRODUCTS IN OVER 60 COUNTRIES ON 6 CONTINENTS WITH MORE THAN 7,000 POINTS OF SALE. THE JEWELRY GIANT CONSTANTLY EXPANDS ITS GLOBAL PRESENCE.

The Challenge.

Pandora needed to find a solution that would help to:

• - Reduce latency and increase data accuracy
• - Increase scalability
• - Eliminate data duplication and usage of extra memory and processing resources
• - Reduce maintenance costs and complexity
• - Make the efficiency and correctness of data exchange more trackable
• - Eliminate operational and analytical data discrepancies.

The Solution.

Together with the client, we redesigned Pandora’s analytics and integration platform from a batch point-to-point solution to near real-time events that can be reused. We’ve built a scalable integration hub based on Event-Driven Architecture, with extended infrastructure around it to meet Pandora's specific needs.

The platform is characterized by high scalability, low cost of maintenance, and high fault tolerance. The product created a digital foundation for integrating all business processes and operating in a digital manner end-to-end. It helps our partner to:

• - Ensure seamless integration between multiple Pandora’s systems, making data exchange between them fast, efficient, and accurate
• - Decrease latency from 36 hours to milliseconds and improve data accuracy
• - Unlock the value of real-time data analytics and provide visibility into all business process
• - Build a solid foundation for Advanced Data Analytics and AI
• - Provide customers with an omnichannel user experience
• - Increase fault-tolerance
• - Enable more efficient logistics and transportation management
• - Ensure future-proof scalability
• - Reduce maintenance costs and complexity

Read more about the project details and Zoolatech Approach at our website: zoolatech.com/retail-digital-transformation-pandora-case-study.html

Client Challenge:
Quickly scale an engineering team that can master a complex lending model, integrate with the in-house org, and take ownership to launch a new mortgages vertical while improving existing products. 

Solution:
Embedded a Kyiv-based team that took full ownership of tech solutions, production support, and hiring. Delivered a microservices architecture, set up Kubernetes + CI/CD + monitoring, implemented a robust rules engine, and executed extensive lender/partner integrations across mortgages and personal loans. Added AWS-based auth/SSO optimization, custom A/B testing, Salesforce back-office integrations, and partner widgets/SDKs. 

Results:

• 7-year partnership with a US FinTech marketplace (7 years, 28 experts).

• Built a mortgage vertical from scratch (rules engine, dashboard, integrations with lenders, >15 integrations).

• Optimized personal loans vertical with microservices refactoring, DB & performance optimization, >20 partner integrations.

• Implemented microservices-based infrastructure on Kubernetes, with CI/CD, monitoring, logging, security & compliance.

Client Challenge:
Scale mobile development fast, handle massive traffic spikes during sales without downtime, and maintain code quality/compliance while shipping continuously. 

Solution:

• Built/optimized key revenue features: multi-payment (PayPal, AfterPay, Apple Pay), guest checkout, loyalty “Rewards,” purchase history, push notifications (price drop, back-in-stock), visual search, iOS widgets, and a dynamic homepage configurable without app updates. 

• Established a 60+ person dedicated mobile org with shared ownership, standardized patterns, proactive prototyping, and 24/7 production support.

• Implemented mobile QA from scratch, secure coding practices, continuous monitoring, and tiered on-call to keep apps stable at peak loads.

Results:

• 10M+ downloads across iOS & Android; ~179k monthly downloads.

• 2× faster development velocity; reliable bi-weekly releases.

• +40% engagement time; +22% purchases; +12% Android revenue lift from guest checkout; improved scalability with no peak-time downtime

Client Challenge:
Redesign a critical, high-throughput service to be more scalable, reliable, and cost-efficient—while eliminating dependencies on deprecated systems and performing a no-downtime cutover. 

Solution:

• Analyzed the legacy OINK (Order-Invoice) service and designed OINK v2 with loose coupling and simplified upstream dependencies.

• Implemented a Kafka-first architecture with the right partitioning for pod scaling; added DLQ across components to avoid missed events.

• Ran v2 in production shadow mode; built a comparator to match v1 vs. v2 outputs; orchestrated safe traffic switch from v1 to v2 in collaboration with OMS/Payments/Tax/ERTM/Sales Audit teams. 

Results:

• 4x reduction in cloud infrastructure costs for a critical high-load service
• 7x faster event processing (from 21k to 150k RTCIM events per hour on 12 pods)
• 4x+ lower memory consumption with the same number of pods
• 3 pods of the new service = 12 pods of the legacy service in terms of throughput
• Zero downtime migration from legacy to the new version
• Eliminated dependencies on services planned for deprecation and improved reliability

Client Challenge:
Siloed legacy systems caused chaotic, inconsistent reporting, poor data quality (e.g., inventory), and slow decision-making. The company needed a modern analytics backbone for reliable, real-time insights. 

Solution:

• Consolidated fragmented sources into a single, event-driven platform (Kafka) with Java microservices and Airflow-orchestrated processing.

• Centralized warehousing (Teradata → migrating to BigQuery) as the single source of truth; exposed insights via Tableau.

• Enabled ML use cases (route optimization, price forecasting, delivery-time prediction) to drive operations.

Results:

• Streamlined operations and cost reduction through better logistics, pricing, and inventory visibility.

• Inventory accuracy improved from ~60% to 90%+ after platform rollout.

• Faster, data-driven decisions via real-time access and standardized dashboards.

Stanfordmag.org is the result of a major transformation in digital publishing for the Stanford Alumni Association editorial team and STANFORD magazine. This transformation included increasing the frequency in publishing new content, revamping internal workflows and processes and changing the way content is organized and resurfaced over time.

Zoolatech was asked to develop the site with a specific focus on mobile readers, including younger alumni, and to focus on ways to make content easier to consume and share.

We improved full-cycle workflow automation for the editorial staff, which allows them to save team members time by using the pre-publishing mode, the "send for approval" option, and inline commenting. 

Also, we built a robust content management system that allows the team to easily update stories and repromote them on the site, in social channels and in newsletters.

Client Challenge:
Peak-season outages and revenue loss from a legacy payment system (graph DB persistence), fragmented fraud workflows (multiple teams talking to Forter), outdated WorldPay transaction parsing, missing checks for specific refund fraud types, and API abuse/credit-card testing without effective rate limiting. 

Solution:

• Built a new payment service and infra from scratch (microservice handling payment APIs; CloudFormation + GitLab pipelines; third-party integrations incl. Discover and internal gift-card API).

• Consolidated fraud decisioning under a dedicated team; created services integrating with Forter, moved async processing from SQS to Kafka, and decomposed 1 service into 7 microservices by business flow.

• Reworked daily WorldPay file processing (Lambdas → Kafka; then AWS Batch for peak-volume scaling). Implemented a proxy/rate-limiting service between Checkout and Payments to block abusive traffic.

• Added NPR (refund) pipeline: Kafka/SQS + Lambda to exchange decisions with an internal ML team, plus an agent review API/UI with Okta/ServiceNow auth.

Results:

• High-availability payment processing with zero downtime at peak; SLAs consistently met with optimized CPU/memory.

• Fraud ops streamlined; avg. transaction processing time cut from ~12 to ~5 minutes.

• Lower chargeback losses and “good customer insult” rates; reduced Card-Free Shopping fraud via targeted account actions and wallet hardening.

• Future-ready, scalable architecture (microservices + Kafka) and modern CI/CD/IaC foundation

Tech partnership with a construction software comp

Client Challenge:
Find senior engineers to speed up time-to-market, modernize legacy apps, and strengthen security while enabling real-time data streaming and enterprise-grade file access/storage. 

Solution:

• Built a high-seniority, long-tenure team and embedded delivery model.

• Implemented Kafka best practices (configuration, security, DR, SLIs/SLOs), custom monitoring, and incident playbooks; took ownership of critical Redis infrastructure.

• Shipped a secure, auditable file access/storage layer (central authN/Z, file-level auditing, storage independence, S3 integration) and executed Rails/Ruby upgrades with deprecation management.

Results:

• 100% Kafka availability for real-time streaming apps; 3× growth in Kafka usage over two years with no major incidents.

• Stronger security, robustness, and scalability across the platform; faster time-to-market.

• Enterprise-ready file platform enabling ISO 19650 compliance and global expansion to large customers

Client Challenge:
Scale customer support cost-effectively, provide 24/7 multilingual assistance, and speed routine tasks (FAQs, pre-qualification forms) without building custom ML from scratch. Human-only support led to delays, inconsistency, and higher costs. 

Solution:

• Deployed an AI chatbot using OpenAI + LangChain with mid-conversation language switching, FAQ automation, and guided form fill for pre-qualification; added a Slackbot for internal teams.

• Tech stack: React widget (frontend), Python/FastAPI service (backend), MySQL, LangChain preprocessing and guardrails for better context and adherence.

• Addressed AI testing/quality: iterative prompt tuning and robust manual validation to handle LLM variability.

Results:
Lower operational costs, faster and more consistent responses, concurrent handling of many conversations, and improved reliability by removing human fatigue/bias—on a scalable foundation ready for new platforms and services.

Client Challenge:
Merchandising ran on disparate apps and manual Excel workflows (pricing, discounts, product setup), causing data inconsistencies, poor collaboration, and slow processes. The client needed a single, standardized platform to unify operations. 

Solution:

• Evolved the Retail Hub into a unified workspace: onboarded existing/new apps and built core features—real-time notifications, issue resolution, task management, and user/permission management (Okta tokens).

• Implemented microservices on Kubernetes with microfrontends; BFF in Node.js, React UIs; HTTP for sync calls and Kafka events (via Proton) for async flows.

• Set up AWS S3 + Athena for reporting/visualizations; enabled secure internal/VPN use with selective external access through Fastly.

Results:

• Unified merchandising operations in a single cloud-based platform
• Improved collaboration across teams via standardized workflows and tools
• Increased scalability and flexibility thanks to microservices and microfrontends
• Real-time insights through AWS-based reporting and analytics
• Stronger security and access control with centralized user and permission management

Client Challenge:
Create a universal, secure, auditable file access and storage solution that meets stringent enterprise security standards—while enabling migrations without downtime or access issues. 

Solution:

• Assessed the monolith and defined a service-oriented architecture to standardize file operations.

• Implemented centralized authentication/authorization and configurable file-level auditing; took ownership of file transfer to/from Amazon S3; enabled storage independence across projects/teams.

• Produced consistent documentation to simplify future support and changes.

Results:

• Secure, scalable file platform with comprehensive monitoring of file operations and access restricted to authorized users.

• Enabled expansion to major enterprises globally; supported compliance with ISO 19650 and related ISO requirements; improved collaboration for construction stakeholders.

Client Challenge:
Media articles from a third party were too generic and duplicative, making it hard to link content to a specific company and explain reputation metric changes in monthly reports. 

Solution:
Stopped relying on third-party texts and built an in-house AI pipeline: LLM-driven summarization/categorization, vector-based deduplication, and alignment of article categories with metric movements; automated report generation with explanatory excerpts and actionable advice. Stack: Amazon Bedrock (Claude for reports; Titan for summarization & embeddings), LangChain.js (Node.js), PostgreSQL + pgvector. 

Results:

• Improved relevance of articles and text shown on personalized customer dashboards

• More understandable and actionable reports by correlating metric changes with categorized, deduplicated content

• Increased cost-efficiency by leveraging multiple Bedrock models through a unified LangChain-based architecture

THIS ORGANIZATION SERVES OVER 10,000 CUSTOMERS AROUND THE GLOBE, WITH OVER 1.6 MILLION USERS. CURRENTLY, THE COMPANY’S DATA HAS TOTALED OVER 500M FILES, EQUATING TO 6 PETABYTES OF STORAGE IN S3.

The Challenge.

The Zoolatech team was tasked with building a solution to secure access to file storage for one of the top construction management SaaS organizations in the United States.

The task was to build a secure, auditable, and scalable solution.

The Solution.

• - Implemented standard access patterns when working with internal files by avoiding insecure and disorganized file path classifications;
• - Embedded authentication and authorization for all systems and users that manage files in order to leverage secure access control;
• - Applied consistent auditing functionalities at the file level;
• - Ensured storage independence. This allows each team and project to distinguish stored files and the associated auditing logs.
• - Additional functionality of the platform includes:
• - Centralized authentication which allows certain levels of customization as per the needs of the project;
• - Configurable auditing at the file level;
• - Increased scalability and functional independence through an implemented microservice architecture;
• - Detailed and consistent documentation, allowing for successful solution updates and the prevention of potential risks of failure.

The Outcomes.

Many potential clients held secure authentication and authorization as mandatory conditions. Once implemented, the new features allowed the Company to onboard many new customers.

This new solution has enabled the Company’s internal teams, stakeholders, and partners to collaborate on construction projects while simultaneously having the ability to safely share access to various documents between all types of IoT devices.

“Zoolatech is the biggest and the most impactful off-shore team that helps to drive our business globally..”

Matthew Paterson,

Technology Sourcing Manager

Client Challenge :
Replace a legacy PHP/Laravel marketplace with a scalable Salesforce Commerce platform, keep operations running (no downtime), and overcome several Salesforce product limits without ballooning costs or timelines. 

Solution:

• Finance & tax automation: Bill.com + Avalara for invoices/bills/taxes & monthly reporting.

• Shipping: Custom tracker integrated with delivery aggregators, near-real-time updates/notifications. Built in ~2 weeks.

• Manufacturer approvals: Secure tokenized 1-click approval directly from email (no login).

• Account matching & self-service payments: Auto link users to business accounts; early invoice payments (prototype in 1 day, live in 4).

• Performance & media: Offloaded images to AWS CDN with smart caching to bypass SFCC limits.

• Punch-out & shelf-life: Procurement punch-out (incl. Shopify-sourced inventory); shelf-life visibility on product pages.

• Data migration: Automated pipelines for customers, manufacturers, orders, products—full integrity and account mapping.

Results:

• Zero-downtime cutover; seamless continuity.

• First major accounting upgrade in 2 weeks of onboarding.

• Up to 5× faster time-to-market and better cost-efficiency vs. vendor estimates.

• $2,000+/month saved by removing outsourced accounting; 1–2 hours/day saved via account-matching automation.

• 2× faster deployments with CI/CD; improved scalability, resilience, and UX.

COGNITOPS IS A B2B SAAS COMPANY HEADQUARTERED IN AUSTIN, TEXAS, AND PROVIDES SUPPLY CHAIN SOLUTIONS TO FORTUNE 500 COMPANIES.

The Challenge.

As the business was growing and scaling exponentially, the data flow was increasing daily, and new challenges were arising:

• - Scale the system to support the exponential SaaS customer base growth, and speed up the onboarding process for new customers.
• - Ensure that increased load or data quality issue for some customers is not affecting others.
• - Ensure bulletproof data privacy and client experience compartmentalization.
• - Ensure that the complexity of the solution and maintenance cost would both remain under control as the number of customers increased and collected data and customizations grew exponentially.

The Solution.

The team created a unique architecture for collecting and storing petabytes of data which allowed analyzing it to generate meaningful insights and statistics. Customers were isolated from each other while the code and core functions were shared. Whenever a new feature was added for one customer, it was possible to reuse it for others. At the same time, the design guarantees data privacy and security for each customer.

Finally, the newly-created architecture allowed smooth onboarding of new customers regardless of the data structures and business processes they are using.

The Outcomes.

By creating a solution that ensures an extreme level of data privacy and security, the team leveraged the platform’s anonymized knowledge for the mutual benefit of all the platform clients.

“We’ve been impressed with the quality of teammates and the speed at which Zoolatech has been able to source and land engineers that could help our internal team. On top of that, they’ve ensured that they choose teammates that have experience with the technology stack we’re dealing with”

Alex Ramirez, 

CEO, CognitOps

COMPLETE SOLARIA OFFERS RESIDENTIAL SOLAR DESIGN SERVICES. A ROBUST SYSTEM THAT MANAGES ALL THE BACKEND PROCESSES ALLOWS THEM TO PROVIDE THEIR CUSTOMERS WITH A DIGITAL EXPERIENCE.

The Challenge.

The Company had to scale quickly due to a market boom in renewable energy.

Moreover, to keep up with a developing technology industry, this Company had to ensure its business strategy was executed with consistency, resilience, and security. Thus, the existing solution required a professional assessment to fully recognize the necessary improvements.

The Solution.

• Engineers from Zoolatech evaluated the situation and working with the Partner's team took the following steps:
• Created a list of improvements for front-end, back-end, and architectural infrastructure.
• Designed the technical roadmap accordingly and blended it into the product backlog.
• Resolved all major technical issues and updated the key technologies to their latest versions.
• Successfully migrated the solution to the AWS cloud.

After fixing the immediate problems and taking technical constraints into account, teams established the Agile process and started to deliver value in a consistent and reliable way. Now the solution runs smoothly, and business features are deployed to production on a regular basis, one sprint at a time.

The Outcomes.

The team significantly improved the system delivery speed, resilience, and scalability by implementing software development best practices and by systematically resolving the technical debt.

As a result, the company can devote more resources to other strategic concerns, like implementing game-changing initiatives.

“It’s not easy to find resources with that level of expertise”

Aida Youssef,

Senior Director of Software Engineering

Client Challenge:
Validate a new location-based platform that helps users discover and engage with nearby experiences (including DigiCal—digital + physical—from home during COVID-era constraints) and establish the marketing foundation for launch.

Solution:

• Product research to align on business model, goals, and audience; simultaneous iOS/Android MVP with React Native.

• Social features (create/apply for events, follow, messaging/communication, sharing photos/comments), plus social network integrations.

• Admin & Business portal, targeted/retargeted advertising engine, analytics tools, and cross-platform maps tuned for both iOS and Android.

• Public landing page to support pre-launch marketing.

Results:
Released the landing page and successfully shipped the MVP to focus groups with a modern, attractive UI. The app enabled event creation and discovery, social interactions, and a robust backend/analytics stack that helps businesses drive foot traffic and sales.

RECON FOOD WAS FOUNDED BY A PRIVATE INVESTOR SPENCER RASCOFF AND HIS DAUGHTER SOPHIA RASCOFF. IT IS A VERTICAL SOCIAL NETWORK TO RESEARCH FOOD PLACES, RECOMMEND AND RECONNECT WITH FRIENDS THROUGH A COMMON LOVE OF FOOD.

The Challenge.

The challenge was to design and create an MVP version of a vertical social network from scratch.

The MVP version had to have a scalable infrastructure and a codebase designed for further development in case the MVP proved to be successful and able to pivot. We had to implement the project within time and budgetary constraints.

The Solution.

• - We had several working sessions with stakeholders, including business owners and investors. The discussions addressed product design, main functionality, customer personas, metrics, potential roadmap, tools, and services required for integration.
• - We prepared a proposal to ensure everyone was on board with the implementation strategy
• - We compiled a team of product-oriented engineers, designers, and product managers to build the application MVP
• - We set up the process and fostered communication to validate the product idea fast and with a cost-efficient approach.
• - We built scalable infrastructure and application architecture to support growing new requirements.
• - We managed the whole development, release, and support process.

The Outcomes.

We managed the whole development process in its entirety from the start and end points to ensure our investors would only have to think of marketing strategy and global product ideation strategies.

After the MVP phase, we proceeded with product ideation and development by participating in user sessions to get insight into the app's functioning. We made data-supported decisions based on the insight gathered from the user sessions.

“The team at ZoolaTech is incredibly collaborative, and we work as a team despite being thousands of miles away from each other.”

Spencer Rascoff

Client Challenge:
Modernize a monolithic customer communication system to boost productivity—respond to business needs faster, add new channels/services, improve CI/CD, consolidate business logic, and reduce operational risk and cost. 

Solution:

• Transitioned to an event-driven, microservices architecture; integrated business + communication events and introduced priority-based triggers (e.g., MFA/verification).

• Implemented Kubernetes with autoscaling and IaC best practices; standardized CI/CD aligned to engineering standards (incl. Confluent Cloud).

Results:

• 2× faster deployment of new features.

• 30% cost reduction by running on K8s pods vs. EC2.

• 6× faster scaling speed and improved service security.

We helped a leading US luxury fashion retailer modernize a Tier-1 customer communication platform by transforming its monolithic architecture into an event-driven, microservices-based solution on AWS. Our team introduced IaC, standardized CI/CD pipelines, migrated workloads from EC2 to Kubernetes, and decoupled communication channels to support new services and faster business change. As a result, the client now deploys new features 2× faster, scales 6× quicker, and has reduced infrastructure costs by around 30% while improving security and reliability.

Client Challenge:
No formal QA process or scalable automation—limited coverage (~300 tests), slow release cycles, and gaps in analytics and localization quality. 

Solution:

• Embedded QA across 7 cross-functional teams; created test backlogs, readable cases, and scalable architectures.

• CI/CD with GitLab; thousands of automated tests run in <15 minutes; Allure for observability.

• Custom analytics interceptors to catch missing/misfired events; AI-powered visual localization validation across markets/devices.

Results:

• ~2,000 automated tests, 75% automation coverage.

• Regression time cut from 48h to 8h; full suite runtime cut from 2h to 17m (parallelization).

• Zero tracking gaps at launch; 95% faster localization QA, zero critical localization defects in prod

Client Challenge:
Rapidly expand engineering capacity to speed modernization and launch new apps, while handling extreme traffic spikes and maintaining quality/compliance across a large retail footprint. 

Solution:

• Established a long-term ODC (2018–ongoing) with 40 ZoolaTech teams covering mobile, cloud, data, DevOps, and high-load systems.

• Implemented event-driven analytics hub, POS data pipelines, QA automation, and privacy tooling (PII tokenization).

• Standardized delivery (CI/CD, observability, release mgmt) and introduced ownership-driven team structure for faster, safer shipping.

Results:

• Time-to-market improved by ~2×; stable peak-event performance; better UX/engagement.

• ODC impact: flexible scale, faster problem-solving, and substantial savings in software/IT ops.

• Scale & reach: 186 experts, 40 teams; mobile apps hit 10M+ downloads with ~179K monthly downloads.