What are the responsibilities and job description for the Applied AI Researcher position at AHUM AI?
Job Title: Applied AI Researcher
Location: Redwood City, CA (On-site)
Job Type: Full-Time
We are a stealth-mode startup founded by a tight-knit team of seasoned engineers and scientists dedicated to solving complex, high-impact problems. Our culture is rooted in trust, autonomy, and a shared conviction that the best innovations stem from diverse perspectives and fearless experimentation. We value deep technical rigor, rapid iteration, and the intellectual courage to challenge the status quo.
As an Applied AI Researcher, you will be a foundational member of our technical team, driving our core model strategy and algorithmic innovation. We are looking for a brilliant applied researcher who thrives on taking cutting-edge AI concepts and translating them into production-ready, state-of-the-art models. You will own the end-to-end lifecycle- from formalizing hypotheses and exploring literature to researching, building, and shipping novel architectures tailored for highly optimized, real-world reasoning.
While we are keeping our exact product under wraps, your work will directly drive our core capabilities. You will take true ownership of several high-impact research areas that we are currently investing in, including:
- Researching and designing highly capable small language models (SLMs) optimized for compute-constrained execution. You will explore novel architectural patterns and push state-of-the-art compression paradigms- such as advanced knowledge distillation and algorithmic quantization- all the way into production.
- Developing and shipping novel methodologies to compress complex reasoning and autonomous agentic behaviors into compact frameworks, rigorously balancing computational footprint against emergent real-world capabilities.
- Leading the applied design of innovative synthetic data generation pipelines to bootstrap training. You will also architect and implement automated, adversarial evaluation frameworks to test model resilience, reasoning depth, and edge-case behaviors.
- Researching and implementing robust, lightweight models capable of performing real-time anomaly detection. You will build systems that parse complex telemetry and system states on the edge to diagnose issues autonomously.
- An advanced degree (MS or PhD) in Computer Science, Mathematics, Physics, or a related highly quantitative discipline with a heavy focus on machine learning or equivalent, battle-tested industry experience.
- A proven track record of researching, training, fine-tuning, and deploying modern LLMs/SLMs. You have deep intuition for LLM training dynamics, objective functions, optimization landscapes.
- Deep familiarity with modern ML ecosystems (e.g., PyTorch, TensorFlow) to implement and ship production-grade, resource-efficient architectures, applying practical principles of model compression, knowledge distillation, and transfer learning.
- Highly comfortable owning the end-to-end pipeline for open-ended problems- from initial idea to shipped model- by designing controlled experiments, utilizing ablation studies, and pivoting based on robust empirical evaluation.
- A strong record of contributions as an author at top-tier AI/ML research venues (e.g., NeurIPS, ICLR, ICML, ACL, EMNLP).
- Experience with causal modeling or uncertainty estimation is a big plus.
We encourage you to apply even if you do not believe you meet every single qualification. Not all strong candidates will meet every requirement as listed. We are looking for high-trajectory problem solvers, relentless curiosity, and a builder’s mindset. We believe this combination is the ultimate recipe for success in our fast-paced environment.
- High Impact: Shape the foundational scientific and product DNA of a zero-to-one product with massive market potential.
- Elite Peer Group: Work shoulder-to-shoulder with a lean, elite team of deep-domain experts who push the boundaries of what is possible.
- True Autonomy: Pursue high-impact applied research avenues with the velocity, freedom, and ownership of an early-stage startup.
- Founding-Level Upside: Competitive compensation, comprehensive benefits, and early-stage equity that reflects your impact as a core team member.