Sam Gao, CEO and co-Founder of DINQ – Interview Series

Sam Gao is a leading AI researcher, engineer and entrepreneur, serving as the CEO and Co-Founder of DINQ, a next-generation talent connection platform for the AI era. Initially trained in civil engineering, he transitioned to AI, publishing over 10 papers at top-tier conferences including NeurIPS, ICML, and CVPR, and contributing to major open-source frameworks such as PyTorch and TensorFlow.

Gao is the second author of DeepFaceLab, the world’s leading open-source face-swapping system, which has earned over 46,000 stars on GitHub and ranked among GitHub’s top ten AI projects in 2020. He also created OutfitAnyone, a universal virtual try-on system recognized among the top 20 projects on HuggingFace Spaces in 2024 and deployed commercially on Taobao, generating over 100 million RMB in annual revenue. Additionally, he authored the Eliza OS AI Agent Whitepaper, a widely cited framework for a Decentralized Trading Agent.

With a global perspective on AI innovation, Gao has engaged extensively with leading researchers, founders and industry pioneers, traveling to hubs including Silicon Valley, New York, Denver, Davos, Singapore and Kyoto. Gao founded the Qingke AI Community, which has grown to over 30,000 public followers and 5,000 experts, offering frontier technology talks, private workshops and networking opportunities. The community is now recognized as one of the most professional and influential networks for researchers who work for xAI, OpenAI, DeepMind, Qwen, Deepseek and more.

You spent several years working on computer vision and graphics for AR and VR at Alibaba Cloud, and later advised on AI-driven proof-of-human systems in blockchain. What personal frustration or inflection point led you to step away from those roles and co-found DINQ?

During my years at Alibaba Damo Academy, I saw cutting-edge tech reach millions of users. However, my biggest frustration wasn’t a technical bottleneck; it was talent misalignment. I saw brilliant PhDs struggle with real-world deployment, while self-taught “coding wizards” were ignored because they lacked a prestigious label. Later, advising on blockchain identity systems taught me the power of “Proof-of-Human.” DINQ is the intersection of these experiences: a mission to provide a definitive, objective Proof-of-Value for anyone building in the AI era.

DINQ is launching at a moment when AI models and compute capacity are scaling faster than the talent required to build and deploy them. From your perspective, what is fundamentally broken about how AI talent is discovered and evaluated today?

The fundamental flaw is “Evaluation Lag.” While AI capabilities scale by the month, hiring remains stuck in a decade-old paradigm:

Keyword Obsolescence: Traditional filters can’t distinguish between someone who simply “uses” ChatGPT and someone who can architect a multi-agent workflow.

The “Pedigree” Trap: Relying on elite degrees or “Big Tech” titles is a lazy proxy for competence. It overlooks the vast ocean of “hidden gems” that are driving innovation in open-source or niche verticals.

Static vs. Fluid: A resume is a snapshot of the past; AI contribution is a living, breathing stream of data across GitHub, Hugging Face, and collaborative platforms.

You’ve described DINQ as a response to the limits of resumes, LinkedIn profiles, and keyword-based hiring. What critical signals about AI researchers and developers are being missed by traditional recruiting systems?

Standard recruiting misses the “behavioral DNA” of a builder:

Iterative Resilience: How does a user refine a prompt or a model until it works?

Contextual Mastery: The ability to bridge the gap between a raw AI tool and a specific business solution.

The “Learning Rate”: In a field where knowledge depreciates every six months, the speed at which someone masters a new framework (like moving from RAG to Agentic workflows) is more important than their total years of experience.

The DINQ Card aggregates code, publications, projects, and collaborations into a single, verified profile. How does this shift the definition of “impact” for early-career AI researchers who may not yet have big titles or well-known affiliations?

The DINQ Card shifts the definition of success from “Who you work for” to “What you’ve actually built.” For early-career builders or non-traditional creators, this is a game-changer. It aggregates verified contributions, whether it’s a high-performing LoRA, a viral AI-generated project, or a critical AI infra bug fix, into a Reputation. It allows a student in a remote area to command the same respect as a Silicon Valley engineer based on the sheer merit of their “Verified Impact.”

On the hiring side, DINQ introduces AI-native search and reasoning rather than static filters. How does this change how companies identify candidates for highly specialized domains like reinforcement learning or multi-agent systems?

Traditional search is binary (Yes/No). DINQ’s search is Reasoning-based. If a company needs someone for “AI agents,” DINQ doesn’t just look for the keyword. It analyzes the candidate’s actual output: Did they solve complex reasoning loops and contribute to Langchain or Dify? How did they handle API latency in their projects? This allows companies to identify “Specialized Generalists”: people with the deep intuition to navigate specific AI challenges that haven’t even been turned into job titles yet.

Having worked inside large platforms like Alibaba Cloud, what do you think big organizations misunderstand most about evaluating real AI capability versus surface-level credentials?

Big organizations often mistake “Past Pedigree” for “Future Adaptability.” They assume that success in a structured, legacy environment translates to success in the “Wild West” of AI. The truth is that AI capability today is about Agency, the ability to take an ambiguous problem and use AI to solve it end-to-end. Large platforms often miss the “scrappy innovators” who are actually moving the needle.

DINQ surfaces collaboration patterns and long-term research trajectory across platforms rather than focusing on isolated achievements. Why is this longitudinal view becoming more important as AI research becomes more interdisciplinary and team-driven?

Innovation is no longer a solo sport; it’s a Collaborative Evolution. By looking at a person’s trajectory across platforms over time, we see their Strategic Consistency. Are they just jumping on every hype cycle, or are they building a deep, interdisciplinary stack? As AI becomes team-driven, seeing how a person interacts with others’ code and research becomes the ultimate predictor of their “Culture Add” and technical leadership.

There’s growing concern that AI hiring is biased toward visibility rather than merit. How does DINQ aim to surface high-impact talent that might otherwise remain hidden or overlooked?

Hiring today favors the loudest voices on social media, not necessarily the most talented. DINQ acts like a “Quantitative Fund for Talent.” We strip away the noise and look at Value Density. By surfacing high-impact contributors who may be “quiet builders” on GitHub, Huggingface or specialized forums, we ensure that merit, not marketing, dictates who gets the best opportunities.

As someone who has operated at the intersection of AI infrastructure, applied research, and now talent systems, how do you see the relationship between AI compute expansion and human expertise evolving over the next few years?

As compute scales, the “Human-in-the-loop” evolves from a doer to an architect. We are moving toward a world where “Expertise” is defined by your ability to steer massive compute resources toward meaningful outcomes. The relationship isn’t competitive; it’s symbiotic. The “AI-Enabled Human” will be the most valuable asset in the global economy, individuals who can orchestrate models, verify truth, and inject creative intuition where algorithms hit a wall.

Looking beyond the January launch, what does success look like for DINQ in reshaping how the AI ecosystem recognizes, develops, and deploys human talent at scale?

Success for DINQ means building the “Trust Layer” of the AI Economy. We want to see a world where a DINQ Card is the only “Resume” you ever need. By 2026, our goal is to have reshaped the global labor market into a true Meritocracy at Scale, where talent is discovered instantly, verified automatically, and deployed to the world’s most urgent problems regardless of geography or background.

Thank you for the great interview, readers who wish to learn more should visit DINQ.