<div>About Kuku</div><div>Kuku is India's leading mobile-first premium storytelling platform. Its flagship apps include Kuku TV for microdrama, Kuku FM for audiobooks and stories, and Guru for microlearning.</div><div> </div><div>Kuku has grown to 400M+ installs across its apps, spanning 1,000+ cities in India. The platform hosts 20,000+ titles across 7+ languages and has empowered over 50,000 creators.</div><div> </div><div style="text-align: justify;">Founded in 2018 by three IITians and Forbes 30U30 honorees, Kuku is backed by marquee investors including MS Dhoni, Fundamentum, Krafton, Vertex Ventures, Granite Asia, IFC, and other leading global venture capital firms.</div>

Kuku FM

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
<ul>
<li>
LLM &amp; 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.
</li>
<li>
Microdrama Intelligence: Build ML/LLM systems to understand transcripts and extract characters, relationships, scenes, story arcs, dramatic beats, hooks, conflicts, reveals, twists, and cliffhangers.
</li>
<li>
AI Screenwriting: Develop LLM-powered systems that transform story ideas into episode arcs, screenplays, scene breakdowns, dialogue, and structured production-ready scripts.
</li>
<li>
Multimodal AI &amp; Video Generation: Build systems leveraging LLMs, VLMs, audio models, image generation, and video generation models to automate the content creation lifecycle.
</li>
<li>
AI Video Pipeline: Design the intelligence layer that converts screenplays into structured, production-ready prompts and instructions for video generation models such as Seedance.
</li>
<li>
Agentic Systems &amp; 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.
</li>
<li>
Story &amp; Character Consistency: Build systems to maintain consistency of characters, relationships, locations, timelines, and story states across multi-episode microdrama series.
</li>
<li>
Evaluation &amp; Optimization: Design evaluation frameworks for measuring screenplay quality, narrative coherence, character consistency, hook strength, cliffhanger quality, and AI-video generatability.
</li>
<li>
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.
</li>
<li>
Architecture &amp; 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.
</li>
<li>
MLOps &amp; Infrastructure: Design and maintain scalable, reliable, and cost-efficient training, inference, data, and evaluation pipelines for LLM and multimodal AI systems.
</li>
</ul>
Preferred Qualifications
<ul>
<li>Education &amp; 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.</li>
<li>
LLM &amp; 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.
</li>
<li>
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.
</li>
<li>
Generative AI Expertise: Strong understanding of LLMs, transformers, embeddings, RAG, instruction tuning, evaluation, and agentic AI systems.
</li>
<li>
Agentic Systems: Experience building LLM orchestration frameworks using tools such as LangGraph, LangChain, or equivalent technologies.
</li>
<li>
Multimodal AI: Experience working with VLMs, speech/audio models, image generation, video understanding, or video generation models is highly desirable.
</li>
<li>
ML Evaluation: Experience designing evaluation frameworks, human preference datasets, automated evaluation, or model-based evaluation for ML/GenAI systems.
</li>
<li>
Strong ML Fundamentals: Strong understanding of machine learning, deep learning, NLP, transformers, embeddings, and model optimization.
</li>
<li>
Engineering: Strong Python programming and software engineering fundamentals with experience building scalable ML services and pipelines.
</li>
<li>
Data &amp; Infrastructure: Experience with distributed data processing, vector databases, cloud platforms, GPUs, or MLOps infrastructure.
</li>
<li>
Research Awareness: Stay up-to-date with the latest advancements in LLMs, multimodal AI, generative AI, AI agents, and applied machine learning.
</li>
</ul>