Converting Ideas into Actionable Solutions

Kalido is an innovative company co-founded by Sanjay Varma (ex Alibaba #3) and Ashvin Sologar (ex Partner McKinsey). Frustrated by the struggle of traditional companies to keep up with agile competitors, they created Kalido to transform the way organisations find, empower, and deploy talent. Kalido's AI-powered talent marketplace democratises opportunity for all.

Kalido’s powerful AI capabilities have the potential to revolutionise internal recruitment and deployment processes. By leveraging advanced algorithms, the platform matches employees to the right teams, projects, and roles with speed and accuracy. Kalido's intelligent profiling system showcases the range and depth of employee skills, providing accurate competency levels and uncovering the true skills ecosystem within an organisation.

Kalido also aims to empower employees by offering opportunities that align with their talents, lowering attrition rates, reducing recruitment costs, and boosting employee satisfaction.

Background

Retro Rabbit has been fortunate to partner with Kalido, witnessing their remarkable journey from an early startup stage to the successful company they are today. From the inception of their vision to their current thriving state, we have played an important role in shaping their technological landscape and supporting their growth.

In the early days, Kalido was founded by Sanjay and Ashvin fueled by their frustration with traditional companies struggling to keep up with agile competitors, they recognised the urgent need for organisations to transform their approach to talent management. They aimed to democratise professional opportunities and empower individuals through an AI-powered talent marketplace. Thus, Kalido was born, with an ambitious mission at its core.

From the very beginning, we joined forces with Kalido as their trusted technology partner. Collaborating closely with their team, we set out to build a robust and innovative platform that would revolutionise the way talent is discovered, empowered, and deployed. We harnessed the power of AI and machine learning to develop sophisticated algorithms and matching systems that could rapidly connect employees with the right roles.

To support Kalido's growth and ensure scalability, we implemented cutting-edge technologies and architectures. As Kalido expanded its services, we continued to innovate alongside them. We pioneered solutions for real-time chat, incorporating diverse media formats, from text and voice notes to images, videos, and documents. Our custom-built UI and intelligent features like sent/delivered/read receipts and push notifications enhanced the user experience and engagement.

Furthermore, we helped Kalido overcome challenges in mobile development, ensuring seamless compatibility with both iOS and Android platforms. Privacy measures, encrypted data caching, and advanced cache invalidation techniques safeguarded sensitive information and optimised data retrieval. Our partnership extended beyond technology implementation. We actively contributed to the growth of Kalido's solution ecosystem.

Features

User Profile with Privacy Controls

Kalido offers advanced profile customisation, empowering users to create personalised profiles. Adding a user’s name, locations, profile picture, roles, biography, contact details, custom content, interests, sustainable development goals, web and social links. Skill recommendations, competency levels, and tags showcase your expertise. Maintain privacy with customisable visibility settings and even import LinkedIn info on mobile. Choose to be searchable by all users or restrict visibility to your networks. Create a tailored profile, control privacy, and foster meaningful connections.

Chat for Seamless Communication

Instant messaging allows users to send messages to visible users based on network memberships and privacy settings. Share text, images, documents, videos, contact details, and locations. Manage chats with archiving, marking as read/unread. Enjoy multi-person chats with channels (network-owned) and group chats (member-owned). Support for public, private, and hidden channels. React, forward, reply with threads, use emojis, polls, share contacts, and locations for enriched conversations. Connect and collaborate effortlessly with our comprehensive chat feature.

Efficient Organisation Networking

Kalido offers private networks for organisations, allowing flexible membership management through various methods such as email validation, email lists, code authentication, and Active Directory single sign-on. Networks can have multiple administrators who can customise network names, branding, descriptions, URLs, and locations. They can send broadcast messages, add/remove members, and hand over administrative control. Customisable questions can be required for membership management, and community guidelines can be posted. Networks can have sub-networks called Channels, with different access levels and membership management options. Non-members can see relevant network information without revealing individual identities. Mobile users can access a feed of recent activities across all networks. Companies can utilise APIs to automate employee information sharing. Streamline organisation networking with Kalido.

Effortless Networking and Skill Recommendations

The recommendations and introductions feature simplifies networking and skill endorsements. Easily introduce members of a user's networks with suggested text and get notified when introductions start communicating. Request introductions based on shared connections. Provide skill recommendations, customise them, and manage received recommendations. Explore others' recommendations and view users they've recommended. Request skill endorsements from a user's own network and edit existing recommendations.

Challenges Faced

Streamlining Infrastructure for Optimal Efficiency

At the beginning of our relationship Kalido was a small startup with a strong focus on cost sensitivity. To address the cost challenges faced and running on limited resources, it was imperative to design a lean system that could operate on the smallest server instances available.

To incorporate machine learning capabilities, a minimum of 6 x p3.2xlarge instances per region were required. However, to mitigate the financial burden, we took innovative measures to optimise the infrastructure. By implementing modifications that allowed the system to run on spot instances, and leveraging spot.io (now part of NetApp), we achieved an impressive 85-90% cost saving on the AWS bill.

Engineering Communication and Data Efficiency

The existing API posed performance challenges as it operated as a single node.js monolith running coffeescript with leaky websockets. This setup resulted in sluggish response times and limited scalability. To overcome these issues, we implemented a comprehensive solution.

First, we replaced the monolithic architecture with Golang microservices, leveraging the power of gRPC for efficient communication. By adopting gRPC and using streams, we achieved bidirectional communication without relying on slow polling techniques

Furthermore, we introduced a lightweight data sync handshake mechanism. The client informs the server about its current data state, and the server responds by providing the necessary updates. This approach allows the client to download only the required data, reducing app startup times by an impressive 50-80%. Additionally, it enables substantial data savings, optimising resource consumption.

Image Management and Dynamic Resizing

Our client faced the challenge of managing images across multiple platforms and formats and preferred not to store multiple sizes of common images, seeking a solution where the API could resize images on the fly. This approach would enable the frontend to effortlessly adapt to different screen sizes without requiring the browser or client to perform image resizing operations.

To address these requirements, we developed an innovative solution centred around an auto-scalable microservice architecture. This solution seamlessly converts images to the precise size and format required by the frontend. The system leverages a Content Delivery Network (CDN) as a frontend interface, which employs a long cache duration. This cache duration ensures that image conversions are only performed once, minimising the need for repetitive resizing operations. As a result, the solution optimises both efficiency and resource usage.

Enhanced Machine Learning for Contextual Matching

Kalido sought a solution that could intelligently match English phrases denoting services or skills to specific projects requiring resources. The challenge lay in accurately and contextually linking the right service or skill to the corresponding project.

Initially, the solution relied on loose matching within ElasticSearch, utilising match probabilities to guide quality. However, we recognised the need for a more sophisticated approach to achieve precise contextual understanding.

To address this, we implemented a novel solution inspired by ChatGPT-like language models. These models possess contextual comprehension capabilities, enabling them to understand the nuances and relationships within English phrases. For instance, the system recognises the connection between terms like "TensorFlow" and "ML", while distinguishing that "Rust SIMD optimiser" and "C++ SIMD expert" are distinct entities.

By leveraging the power of this advanced Language Model, we significantly improved the accuracy and intelligence of our matching system. The solution now provides a more precise and contextually aware approach to link services and skills to the appropriate projects.

Results Achieved

Achieving Idea to Implementation

We embarked on a collaborative journey with Kalido to transform their visionary idea into a fully implemented reality. From concept to delivery, we worked closely with the Kalido team, crafting a comprehensive roadmap and leveraging cutting-edge technologies to design a robust and intuitive platform. Through iterative development, rigorous testing, and ongoing support, we delivered a scalable, secure, and high-performance solution that aligned perfectly with Kalido's vision. Our collaborative efforts empowered Kalido to realise their goals and achieve remarkable success, showcasing our commitment to delivering transformative software solutions.

Kalido’s Recognition as a 'Tech Pioneer’ for 2020 by WEF

Kalido's significant societal impact and technological advancements led to their esteemed selection as a 'Tech Pioneer 2020' by the World Economic Forum. This prestigious accolade placed Kalido among a distinguished group of start-ups and scale-ups, known as the Global Innovators community, comprising the most promising and pioneering organisations worldwide. The World Economic Forum recognised Kalido's transformative contributions to technology and business model innovation, affirming their position as a trailblazer in their industry.

Successful Launch of Talent-Centric Internal Talent Marketplace

Kalido achieved a significant milestone by launching an enterprise product, an internal talent marketplace with a strong focus on talent-centricity. The initial pilot phase involved collaboration with 12,000 users from HCL. The positive impact and potential of the platform led to the signing of a contract to scale the platform to an impressive 75,000 employees in Digital Foundations, HCL's largest business unit.

About the author

Theo Crous profile picture

Theo Crous

Theo Crous

I co-founded Retro Rabbit in 2003 with my university friends. In my spare time I like to raise my two kids, drink wine, hit the mountain bike trails and learn how to sail with my beautiful wife. Read more from Theo Crous...