Senior AI/ML Engineer

About The Job As a Senior AI/ML Engineer, you will be instrumental in building Kuku’s next-generation AI-powered content creation platform, with a focus on LLMs, instruction tuning, multimodal AI, and AI-generated microdramas. You will work on building systems that understand thousands of hours of existing microdrama content and transform creative ideas into engaging screenplays and production-ready video generation instructions.

This role will involve working across LLM training, content understanding, agentic workflows, screenplay generation, evaluation, and AI video generation, while owning key ML systems end-to-end. You will work closely with content, product, and engineering teams to build AI systems that directly impact how stories are created at Kuku.

Responsibilities

• LLM & Instruction Tuning: Design and build large-scale instruction-tuning datasets from Kuku’s microdrama corpus and fine-tune LLMs for screenplay generation, dialogue, story continuation, rewriting, and content transformation.
• Microdrama Intelligence: Build ML/LLM systems to understand transcripts and extract characters, relationships, scenes, story arcs, dramatic beats, hooks, conflicts, reveals, twists, and cliffhangers.
• AI Screenwriting: Develop LLM-powered systems that transform story ideas into episode arcs, screenplays, scene breakdowns, dialogue, and structured production-ready scripts.
• Multimodal AI & Video Generation: Build systems leveraging LLMs, VLMs, audio models, image generation, and video generation models to automate the content creation lifecycle.
• AI Video Pipeline: Design the intelligence layer that converts screenplays into structured, production-ready prompts and instructions for video generation models such as Seedance.
• Agentic Systems & GenAI: Design and implement advanced agentic workflows using frameworks such as LangGraph to orchestrate story planning, screenplay generation, critique, rewriting, continuity validation, and video prompt generation.
• Story & Character Consistency: Build systems to maintain consistency of characters, relationships, locations, timelines, and story states across multi-episode microdrama series.
• Evaluation & Optimization: Design evaluation frameworks for measuring screenplay quality, narrative coherence, character consistency, hook strength, cliffhanger quality, and AI-video generatability.
• Technical Ownership: Provide technical leadership in ML model formulation, experimentation, architecture, and deployment. Take end-to-end ownership of critical AI systems and guide other ML engineers.
• Architecture & Strategy: Define the technical architecture and roadmap for Kuku’s AI content generation platform, evaluating foundation models, fine-tuning approaches, RAG, preference optimization, and agentic architectures.
• MLOps & Infrastructure: Design and maintain scalable, reliable, and cost-efficient training, inference, data, and evaluation pipelines for LLM and multimodal AI systems.

Preferred Qualifications

• Education & Experience: Bachelor’s or Master’s in Computer Science, Machine Learning, Statistics, or a related engineering field, with 4-6 years of relevant experience, including experience owning and delivering complex ML/AI systems.
• LLM & Model Training: Hands-on experience training or fine-tuning large language models using frameworks such as PyTorch, Hugging Face Transformers, PEFT/LoRA, or equivalent technologies.
• Production Experience: Proven track record of productionising ML/GenAI systems, including designing and managing end-to-end ML systems, data pipelines, model serving, and monitoring.
• Generative AI Expertise: Strong understanding of LLMs, transformers, embeddings, RAG, instruction tuning, evaluation, and agentic AI systems.
• Agentic Systems: Experience building LLM orchestration frameworks using tools such as LangGraph, LangChain, or equivalent technologies.
• Multimodal AI: Experience working with VLMs, speech/audio models, image generation, video understanding, or video generation models is highly desirable.
• ML Evaluation: Experience designing evaluation frameworks, human preference datasets, automated evaluation, or model-based evaluation for ML/GenAI systems.
• Strong ML Fundamentals: Strong understanding of machine learning, deep learning, NLP, transformers, embeddings, and model optimization.
• Engineering: Strong Python programming and software engineering fundamentals with experience building scalable ML services an…