FireworkTV, a decentralized short video network, was facing challenges with its existing machine learning (ML) infrastructure. The ML team recognized the limitations of their current system, which included lagging productivity, growing overhead costs, and a lack of automation. These issues were hindering the performance, quality, and reliability of their video recommendation model. The model, which is crucial for engaging users and driving ad revenue, needed to deliver highly accurate and real-time recommendations based on user-video interactions and specific content features. However, the existing infrastructure, based on Lambda and PyTorch, was not only expensive but also cumbersome, limiting the project's potential to scale and grow. The team sought to build a new, more efficient infrastructure on AWS to drive improvements.

Appen, a leading provider of high-quality training data for AI systems, was facing challenges with its manual ticketing system. The system was inefficient, time-consuming, and prone to errors, particularly in ticket categorization. The Contributor Success team, consisting of only two members, was overwhelmed with over 11,000 tickets per month, leaving them with only seconds to resolve each issue. The manual categorization often resulted in miscategorized tickets, leading to delays in response times and dissatisfied customers. Appen needed a solution to automate its ticketing system, improve support efficiency, reduce ticket handling time, and decrease contributor churn.

Lane Health, a healthcare lending company, was facing challenges with its HSA Advance applications. The applications were developed on a no-code platform, which limited their flexibility and posed ownership and maintenance challenges. The applications lacked a proper versioning system or database rollback mechanisms, making each new product release a risky endeavor. They also lacked tools for testing, logging, monitoring, alerting, and database migration and management. Lane Health wanted to improve the HSA Advance applications by making them more secure, scalable, flexible, reliable, and cost-efficient. They were looking for ways to quickly introduce a more advanced technology stack and infrastructure, and to migrate the applications to the AWS cloud. The team hoped to gain capabilities to innovate faster while achieving more stable releases, reducing the Total Cost of Ownership, and achieving HIPAA compliance to store PHI data. The product had to be launched in less than four months, requiring a reliable partner to get it all done.

GoCheck Kids (GCK), a comprehensive photoscreening and visual acuity app, was seeking to enhance the image classification component of its pediatric photoscreening application through machine learning. The cloud-based system was data-rich, and it was hypothesized that machine learning could improve the app's usability by supplementing image analysis and improving user actions to capture the best image possible. GCK needed a robust and resilient machine learning infrastructure to run experiments on a dataset of over one million images faster and more cost-efficiently, to train a highly accurate image classification model for the application. Additionally, GCK wanted to alert the user in real-time to retake the image when the image was captured while the child was not looking directly at the camera. This required a machine learning infrastructure to prepare data, run experiments on the dataset, and build and train new image analysis and classification algorithms faster and more efficiently.

IMVU, the world’s largest avatar-based social network, was facing challenges with its aging on-premise data platform. The company wanted to enhance and re-architect their platform to support advanced analytics and Machine Learning use cases. However, with an exponentially growing data volume and a monolithic Hadoop architecture, the IMVU team was struggling to efficiently utilize user-generated data. The existing infrastructure limited innovation and capacity for advanced analytics. IMVU’s analysts lacked the tools to rapidly generate business-critical reports on customer in-game behavior at scale. They were working with historical data in batches, which resulted in late reports, inaccurate assumptions about customer in-game purchases, slower sales, and loss of profit. The analytics team also lacked a test environment to efficiently check analytics assumptions. The platform was powered by a 90-node on-premise Hadoop cluster, which was not cost-efficient and resulted in high costs and low efficiency.

InMarket, an omnichannel marketing platform, was grappling with an inefficient legacy data platform that was unable to handle the growing volume of real-time location data collected from multiple sources. The platform, built using 50 AWS nodes and 400 bare metal nodes managed by Apache Mesos, was processing over 5 billion events daily, leading to delays, bottlenecks, and inefficiencies. The platform's performance was subpar, with a job success rate of only 40%, and 60% of Apache Spark jobs were randomly aborted in the system. This inefficiency led to developmental delays, inaccurate timeline projections, and a slow handoff process from data scientists to data engineers and operations. These issues were detrimental to InMarket's ability to attract marquee brands and slowed down revenue growth.

LeadGenius, a marketing automation and demand generation company, was facing significant challenges with its data processing pipeline. The pipeline was inefficient due to a high amount of manual processes, including data incorporation and verification. This inefficiency led to bottlenecks, slowing down data delivery to customers. The data, parsed from various sources, had to be verified carefully, which when done manually, further slowed down the process. The pipeline was also lacking in terms of data quality and data consistency due to the variety of data sources and the reliance on manual processing. The company needed a solution that was not only automated but also fault-tolerant and scalable, capable of running on-demand in case of any issues with its components.

The client, a global healthcare leader based in the United States, was looking to accelerate and scale the adoption of AI/ML across its organization. The company generates substantial amounts of data from various sources, including customer interactions, sales transactions, social media activity, and product usage. However, without a robust Machine Learning Operations (MLOps) platform, it was a challenge for the organization to effectively scale and manage their AI/ML workflows. This resulted in inefficiencies, increased costs, and slower time-to-market for new products. The client’s data scientists and ML engineers were looking for ways to simplify the deployment of AI/ML into production environments, particularly when using MLOps practices and the Amazon SageMaker suite of services. The client was transitioning from legacy infrastructure, but its engineers could not access and discover the unified and integrated workloads quickly and efficiently enough to meet the company’s vision for AI transformation.

PSC Biotech, a global life sciences consultancy, was facing challenges with its document processing operations. The company was looking to enhance its operations by automating its existing pipeline with AI, specifically to process FDA Form 483 observations faster, more accurately, and on a larger scale. The company's existing processes were manual, leading to high processing costs, risks of errors due to human factors, and a throughput rate that depended entirely on the number of employees. The accuracy of document processing was also low and fluctuated significantly over time. Given the sensitive nature of the HCLS business operations, any process that is slow, inefficient, and prone to errors poses a huge risk. PSC Biotech handles thousands of FDA Form 483 observations per year, and the need to automate its document processing pipeline was long overdue. The company expected to decrease time spent on manual review of observations, decrease costs of form processing, mitigate risks of infractions made by mappers and reviewers, increase throughput rate, and increase the accuracy of document processing by adopting AI and implementing an ML-powered Intelligent Document Processing (IDP) solution.

Pr3vent, a Silicon Valley-based diagnostic company, was faced with the challenge of improving patient diagnosis and eye screening availability through computer-aided diagnosis. The company aimed to scale doctors’ expertise through AI, with the goal of reducing the per-screen cost for better accessibility to 4M infants in the US alone while increasing diagnosis accuracy. The challenge was to utilize the power of AI to combat preventable vision loss in infants. Due to the scarcity of trained doctors who can diagnose eye diseases by a newborn’s retina, the team’s vision was to marry Deep Learning and data to scale the expertise of ophthalmologists who can, to cut per-screen cost, increase accuracy, and improve screening availability. The solution needed to be highly accurate in detecting pathology in a newborn’s retina, to receive FDA approval. This required Pr3vent to accurately label a database of 350K fundus and retina images by a team of experienced ophthalmologists, build an AI-driven image analysis and anomaly detection engine, and develop an application for ophthalmologists to handle retina images.

Nitrio, an AI company specializing in sales optimization, was facing significant challenges with its Natural Language Processing (NLP) platform. The platform relied heavily on manual rules and heuristics-based models, which led to bottlenecks and scalability issues, hindering Nitrio's growth. The existing platform was unable to ensure the required level of accuracy for sentiment analysis of rep-to-lead messages, resulting in a significant number of messages being outsourced to a third party for manual analysis. This not only increased service costs but also created further bottlenecks and scalability issues. The platform's infrastructure demonstrated tight coupling between services, increasing their dependencies and negatively impacting team performance, causing data quality and consistency issues. Nitrio's platform was designed to efficiently analyze inbound rep-to-lead messages to extract their intent and collect useful data about every sales representative's performance. However, the reliance on manual processes and the inability to ensure 95% certainty in message intent identification were major setbacks.

Swiftmile, a universal charging platform for micromobility, was preparing for its next phase of global expansion. However, the company's growth potential was limited by its existing platform's monolithic infrastructure and backend logic. The platform could not support more than 600 stations while handling about 30 messages per second, causing substantial time losses when processing transactions beyond this limit. This major bottleneck limited the platform’s performance, scalability, and cost-efficiency. Swiftmile needed a robust, highly scalable, cloud-based platform that could maximize scaling potential, simplify the integration of new communication streams, and prepare for worldwide expansion while keeping down infrastructure costs. Additionally, Swiftmile sought to make their platform data-driven, with the ability to collect, process, and analyze streaming data in real time to improve user experience and explore data business opportunities.

InMarket, an omnichannel marketing platform, was grappling with an inefficient legacy data platform that was unable to handle the growing volume of real-time location data collected from multiple sources. The platform, built using 50 AWS nodes and 400 bare metal nodes managed by Apache Mesos, was processing over 5 billion events daily. However, it was plagued with delays, bottlenecks, and inefficiencies. The platform's job success rate was a mere 40%, with 60% of Apache Spark jobs being randomly aborted in the system. This led to developmental delays, inaccurate timeline projections, and a significant reduction in InMarket's ability to attract marquee brands, thereby slowing down revenue growth. The time taken to hand off a data pipeline from data scientists to data engineers and then to operations for deployment in production was estimated to be up to twelve months, which was unacceptable given InMarket's business model.

Nugs.net, a music streaming and live recording platform, aimed to enhance its user experience by delivering a superior platform for music fans who enjoy live performances. The company wanted to increase user retention, conversion rates, and profitability by improving UX personalization. The challenge was to incorporate AI recommendations to improve content discoverability, listening diversity, catalog observability, and the scope of artist following. Nugs.net also faced growth seasonality, with users typically sticking to newly performed and released concerts and recordings. This was highly dependent on artists’ schedules, which could be disrupted by various factors, making for an unpredictable growth strategy. Another challenge was the user tendency to stick to one or two favorite artists or genres, making Nugs.net even more dependent on growth seasonality for retention.

Second Genome, a biotechnology company, was seeking to accelerate and scale its microbiome drug discovery and development. The company wanted to improve data ingestion and staging, and refine the codebase of its data platform. Operating in a highly regulated pharmaceutical industry, Second Genome needed to enhance data security compliance to create a safe drug research and development environment for its clients and partners. The company was also looking to handle microbiome data more efficiently to speed up microbial research, drug trials, and discovery. As part of the healthcare industry's transformation towards personalized medicine, Second Genome was aiming to identify responder/non-responder populations and determine the optimal approach to therapy. The challenge was to enhance its data platform to make it faster, more scalable, secure, and compliant.

Nexant, a globally recognized software, consulting, and services company, was facing challenges with its on-premises IT infrastructure. As part of a multiyear turnaround effort, the company sought to transform its IT infrastructure and migrate its operations to the cloud. The goal was to enable faster iterations, assess, deploy, and support customer applications more efficiently, speed up development and release cycles, eliminate unnecessary heavy lifting, and scale its entire system for faster growth. However, deploying new resources and maintaining existing ones was time-consuming, slowing down the development and delivery of applications, and causing operational inefficiencies. To overcome these challenges, Nexant partnered with Provectus, an AWS Premier Consulting Partner, to reinvent its IT infrastructure and migrate it to the AWS cloud.

VTS, a commercial real estate leasing and asset management software and data company, aimed to become the decision-making platform for the commercial real estate industry. To achieve this, they wanted to efficiently productionize Machine Learning (ML) models and build new models iteratively using AWS services. Their goal was to accelerate the time to market for ML applications, reduce human errors, and lessen the effort from their Data Science (DS) team. One of the ML models they prototyped was designed to predict leasing deal outcomes. However, they faced challenges in integrating this predictive model into the core user experience. While their data scientists were capable of delivering the model in ad hoc environments, they found it difficult to deploy the model in production using the existing infrastructure of the VTS platform. In essence, VTS had excellent data scientists but lacked the necessary AWS and MLOps expertise to complete the task.

Kahoot!, a platform for creating, sharing, and playing learning games or trivia quizzes, has experienced significant growth since its inception in 2012. With over 550,000 paying users and more than 1.5 billion participating players in 200 countries, the company faced the challenge of effectively managing and utilizing its vast product usage data. Despite having adopted Amplitude, a product intelligence platform, Kahoot! was only using a few functionalities and tracking minimal events. The company was suffering from a classic bottleneck where all data requests had to go through the data analysts. This situation was not sustainable given the company's growth and the increasing need for data-driven decision making across different departments.

TripActions, a corporate travel management organization, was facing a significant challenge with its existing infrastructure and data storage solution. The increasing volume of historical indexed data was straining the company’s infrastructure and primary storage solution, slowing down its performance and causing an ever-increasing cost of ownership. The historical data was never cleaned while analytical data was stored across various databases in different formats, creating multiple data silos and making data unavailable for analytics and machine learning. The company’s existing data solution was failing in terms of analytic capacity and scalability, which increased operational costs, slowed down onboarding of new clients, and stifled business growth. The initial architecture and data solution were based on Amazon Elasticsearch, which proved to be inefficient and expensive when data volumes increased. Data was schemaless, and there was no mechanism to join data from different databases. Partial data in Amazon S3 was stored in JSON format and synced with one-day lag, with no partitioning, which delayed TripActions’ reaction to issues or changes in data.

Canva, an online design and publishing platform, was facing a challenge in managing and utilizing its vast data. The company wanted to empower non-technical stakeholders with self-serve data to explore different areas as needed. They had a data warehouse, but the barriers to entry were too high for the average user. To grow Canva at scale, non-technical people needed to segment audiences and create funnels. It was difficult for product managers to dive into new releases and see how new features performed or get a breakdown of a funnel. Shortly after launch, the team realized the need for a more detailed product analytics solution.

Cancel Timeshare, a Myrtle Beach based company, was experiencing rapid growth and needed a tool to manage and analyze data from Stripe, handle dunning, and consolidate customer information. The company, which helps timeshare owners exit their contracts, was manually tracking revenue across multiple apps and spreadsheets, which was time-consuming and inefficient. Additionally, they were losing revenue due to failed payments, a common issue for subscription-based businesses. The challenge was to find a solution that could streamline their data management, recover failed payments, and support their customer service goals.

Seagate Technology, a renowned data storage company, was facing a significant challenge in transitioning its business model from OEM and channel distribution to a subscription-based model. The company wanted to consolidate its forecasting processes across various teams and business segments, and sought to use deal rooms for collaboration with both internal and external stakeholders. However, Seagate's sales teams lacked the necessary experience in the subscription business model. They also struggled with the quick movement of large volumes of data in the revenue cycle with customers and had an insufficient customer-facing and user sales organization that could interact directly with customers within the Go-To-Market (GTM) strategy.

UXPin, a code-based design tool company, was facing a significant challenge in consolidating and analyzing their subscription data. Initially, they used both PayLane and PayPal to process payments. However, when they moved to the United States, they were unable to migrate their Poland-based PayPal subscriber data. This was a significant issue as these customers constituted a large portion of their revenue. To address this, they decided to maintain their Polish PayPal account while also establishing a separate payment processor in the US. This decision led to the challenge of maintaining two separate payment processors and the need to consolidate and analyze data from both. UXPin built internal tools to analyze activity, but these tools required ongoing maintenance, often crashed, and did not provide the insights they needed. By October 2020, 99% of UXPin’s customers were paying via Stripe, and they needed a solution to efficiently analyze this data and support their growth.

Sync with Connex, a successful SaaS company, faced several challenges in its growth journey. The company needed to track data from Stripe, recover failed credit card payments, and optimize their cancellation flow. They were also struggling with manual data entry, a tedious and error-prone task that consumed significant time and resources. Additionally, they were using PayPal for processing payments, which offered little additional value and lacked the necessary reporting tools. The company was essentially operating with limited visibility into important customer details. The lack of an effective dunning system led to failed payments and a high volume of customer complaints, which further strained their customer success team.

MetricFire, a small yet powerful team, was facing challenges in efficiently analyzing data from Stripe and segmenting their customers. The company, which offers infrastructure, system, and application monitoring using a suite of open-source monitoring tools, had to delegate billing and financial reporting to their Business Operations Manager, Elliot Langston, due to a team reorganization. Langston was using MetricFire’s Stripe account for these tasks. However, Stripe’s in-app metrics feature was limited to just 16 metrics and often gave misleading Monthly Recurring Revenue (MRR) figures due to occasional bugs and counting non-recurring revenue towards MRR. Langston had to create dummy accounts in Stripe to correct these inaccuracies, a process that was not sustainable and risked billing mistakes. Additionally, Stripe did not allow customization of date views and lacked the ability to segment customer data, making it difficult for Langston to efficiently organize customers by geographic location.

Smart Passive Income (SPI) was facing a couple of significant challenges. Firstly, they were using multiple tools to run their business, each with its own analytics dashboard. However, these tools were not interconnected, limiting the amount of usable data SPI could access. This lack of comprehensive data made it difficult for the SPI team to make informed business decisions. Secondly, SPI was struggling with the recovery of failed payments from members of their SPI Pro community. This was causing a loss of revenue and potential disruption in the continuity of memberships.

Branch, a company with a finance team playing a crucial role in their country-wide expansion plans, was struggling with the limitations of spreadsheet-based planning. The team was dealing with complex multi-dimensional modelling due to four product offerings, an expansion from 5 to 50 states, and a need to plan by product, geography, and acquisition channel. The limitations of Excel, including lack of model flexibility, increased workbook loading times, and endless manual transformations, slowed the finance team to the point of paralysis. As a venture-backed company, managing cash runway was crucial for Branch. However, constantly changing requirements, broken cell references, and the lack of an audit trail made the team lose confidence in the quality and accuracy of their models. Furthermore, the size and manual nature of each model made communicating results to decision-makers virtually impossible.

PuroClean, a leading franchise system for emergency property damage remediation, faced a significant challenge in managing their Local SEO initiatives across their 360 franchise locations in the USA and Canada. The marketing team struggled to get all franchise locations on the same page and bought into their structured, consistent Local SEO initiatives. Being a small team, they had to commit additional time to overcome this challenge, which led to other key functions of the marketing team being neglected. They identified the need to improve their understanding and management of NAP (Name, Address, Phone number) in relation to local business listings, but lacked the time to address this. The scale and scope of the challenge were largely unknown due to the novelty of Local SEO efforts to the business.

Reboot Online, a London-based digital marketing agency, faced a significant challenge in managing the local SEO needs of their clients. They noticed that many of their clients had inconsistent citation profiles, including outdated addresses and phone numbers, which confused search engines about the actual identity and location of these businesses. The team tried several methods to clean up these citation profiles and expand their clients' online footprints, including other citation services and manual submissions using Excel spreadsheets. However, these methods proved to be slow, time-consuming, and ineffective. Reboot Online needed a solution that could deliver a quality service at scale, without requiring a lot of time to implement, and still allowed them to maintain control.

My Social Practice, a digital marketing agency for dental practitioners and other medical professionals, faced a significant challenge in helping their clients understand the intricacies of digital marketing strategies and tactics. The agency provides dental marketing services, training, and support to thousands of dental practices, many of which lack a comprehensive understanding of digital marketing. The complexity of explaining local SEO rankings, Google Business Profile workings, Google Maps ranking process, location factors, and the search algorithm often led to confusion among clients. The agency needed to provide an SEO report that was simple to understand to prevent client confusion and cancellations.