Projects

Growing-Mate aims to leverage advanced AI capabilities along with innovative methods and strategies to provide clients with efficient, intelligent, and personalized marketing solutions. We are committed to developing smart marketing tools and platforms to enhance marketing efficiency and effectiveness. Optimizing user acquisition and operational costs through AI-driven growth strategies Our goal is to develop AI models that learn from multi-modal user data to automate the entire user growth funnel (AARRR: Acquisition, Activation, Retention, Referral, Revenue) operations management. This is achieved through: AI algorithms that provide deep insights into user needs Identification of external trending topics (with multi-modal level understanding) Generation of high-quality content with high exposure potential and understanding of the user’s promotion targets Implementation of omnichannel distribution The simple structure is shown below: Our distribution strategy extends beyond platform content to include multi-modal advertisements, viral social media posts, peripheral marketing accounts, and customer service responses. The synergy of these four elements creates a flywheel effect, continually driving growth. By leveraging advanced AI capabilities, G-Mate aims to revolutionize user growth strategies, making them more efficient, targeted, and cost-effective for businesses. Our vision is to develop a co-pilot for Key Opinion Leaders (KOLs), guiding them from content creation to monetization. We aim to achieve this by leveraging the integration capabilities of tools like Luban. Unless absolutely necessary, we will avoid developing features in-house; instead, we’ll rely on stable APIs provided by major platforms. For the Minimum Viable Product (MVP), our initial focus will be on the content creation phase. The first strategy is to replicate trending content (case studies). The second strategy is to emulate successful KOL accounts (case studies). Value Proposition: The Importance of Emerging KOLs Emerging KOLs hold more value than established ones because they are favored by exposure algorithms. Established KOLs benefit from platform-weighted support, making it challenging for smaller KOLs to replicate their success. It’s akin to trying to replicate the strategy of a major figure like Duan Yongping, which is nearly impossible; however, replicating the approach of someone like Mr. Chen is feasible. Similarly, KOLs can only replicate the success of those within their own follower base. Target Audience: KOLs Below the Mid-Tier Level Our target audience is KOLs below the mid-tier level, essentially individual influencers who function as their own mini-MCN agencies. For these KOLs, the cost of monitoring social media trends, sourcing creative materials, analyzing results, monetizing content, and performing data analysis is prohibitively high. With our quantification tools, they can compete against larger agencies while significantly reducing the cost of trial and error. By employing reverse content selection, we help KOLs determine whether their content style resonates with an audience. KOL Persona as Strategy Establishing a persona is the strategy for KOLs. This involves setting up a persona for the KOL, then subscribing to similar large influencers or trending topics with high relevance. Capturing Trade Signals Traffic Radar: Present trending topics through swipe gestures or a waterfall format. Similar to platforms like Sinan Planet, we can offer trend subscriptions categorized by KOL persona, region, social media platform, and custom tags across three tabs. Content Selection & Historical Performance Analysis Once a trending topic is selected, the platform provides the source URL and historical data analysis for similar content, along with pre-drafted derivative content. The system will also assess the alignment with the KOL’s pre-established persona, offering exposure estimates, enhancement strategies (e.g., whether to boost content using tools like Dou+), and product placement recommendations. Fine-Tuning the Trade Plan Users can modify AI-generated derivative content through chat or text modes, providing prompt tags as needed. Content Advisor: Risk Management & Enhancement Our content advisor offers risk management (e.g., flagging sensitive words, political direction, persona integrity) and content refinement suggestions based on high-exposure examples. Additionally, it provides recommendations on optimal posting times and relevant tags. Portfolio Management Compare actual exposure with expectations, and decide whether to boost content using Dou+ (increase exposure) or pivot to new, similar content for another attempt (mitigate loss, reinforce position). Monetization The platform extends into monetization by offering strategies to increase follower retention through extended reading and product recommendations. Content Replication Users can immediately repost the same content on similar accounts, different types of accounts, or even across platforms (e.g., converting a text post into an image post for Instagram). Market Manipulation (Advanced Strategy) We offer tools that allow KOLs to exploit information asymmetry. For example, after promoting a “Golden Cross” strategy, a KOL might generate dependency on the tool among their audience. By simultaneously providing tools for retail users to discover similar opportunities, they can manipulate the market. Using their own capital (bot networks), they can drive up the “stock price” (create a trending topic), signaling others to follow suit, ultimately turning it into a “GameStop” scenario. We also propose a crowdsourcing platform to launch tasks such as follower acquisition, likes, and engagement, transforming the GM bot network into a semi-official monetization channel. Information asymmetry is key—both in timing and platform selection. Thus, beyond rapidly capturing trends, cross-platform content repurposing is another viable strategy, like pairing a trending tweet with music and reposting it on Twitter (X) and TikTok.

Aug 10, 2024

To create an AI-powered virtual HR employee, primarily serving small and medium-sized enterprises (SMEs). We aim to empower and simplify the recruitment process, enabling companies to complete hiring tasks efficiently with minimal or even zero human resources. This solution addresses the critical challenges of recruitment difficulties and high costs faced by SMEs, tackling a fundamental issue that impacts their survival and growth. Screenshots of the product:

Apr 20, 2024

An accelerator and collaboration platform for AI application development. Product link: Lupan Provide developers with a cost-effective AI tool integration solution by offering a one-stop service for finding, integrating, and debugging tools. Achieve the goal of using natural language to issue tool usage instructions to large language models (LLMs) and interact with them. This is accomplished through a unified data standard encompassing APIs, LLMs, and UI. Foster a co-creation ecosystem through community operations, aiming to establish a crowdsourced model cycle for AI tool data. By streamlining the development process, Lupan empowers developers to innovate more efficiently in the AI application space.

Feb 26, 2024

Watt is a multi-agent task compiler that connects various real-world APIs. By fostering an executor (workflow automation tool) and a multi-agent orchestrator, it empowers agents to work together seamlessly, tackling complex tasks.

Nov 26, 2023

The project aims to develop and implement a high speed stereo vision system and apply it onto unmanned ground vehicles (UGV) with multi-sensor fusion with deep neural network. The method uses bird’s-eye view representation space to preserve both geometric and semantic information. Project Page: https://ugv_stereo.gitlab.io/ Some video demos are available on Google Drive. The project consists of three main parts: visual inertial SLAM, 3D reconstruction and tracking, and lane detection. Visual inertial SLAM is used to estimate the pose of the camera and the 3D map of the environment. The 3D reconstruction and tracking part is based on the 3D reconstruction algorithm, which is used to reconstruct the 3D scene. The lane detection part is based on the lane detection algorithm, which is used to detect the lane. The project is implemented in C++ and ROS.

May 20, 2019