AI Application Engineer

AI Applications Engineer – Annotation Tooling and Data OperationsAbout KGeN/KAIKGeN’s “KAI – Training & Evaluation” business taps a network of 1M+ verified experts across 60+ countries and 20+ languages, delivering specialised datasets across domains such as coding, STEM, healthcare, content, design, legal and general data collection and labeling power large‑language and multimodal AI systems.KGeN is one the fastest growing startups with revenue growth to $50M+ ARR within 3 years. Funded by the most marquee investors - Accel, Prosus, Sequoia, Nexus and Lightspeed.Role OverviewYou will design, develop, and deploy intelligent tooling and systems that enhance the efficiency of KAI’s GenAI operations. Your work will focus on automating human-in-the-loop (HITL) pipelines, improving data labeling and evaluation workflows, and leveraging AI/ML to optimize annotation quality, routing, and workforce productivity. This role is ideal for engineers passionate about combining applied ML, backend systems, and data infrastructure to improve real-world AI training at scale.Key Responsibilities1. Build the Multimodal AI Annotation PlatformArchitect and develop a modular annotation system supporting images, audio, video, and text.Integrate open-source frameworks (Label Studio, CVAT, Diffgram, Whisper, Pyannote, MediaPipe, Rekognition, Vertex AI, etc.) with custom logic for quality control, automation, and AI-assist.Implement APIs and SDKs to manage data ingestion (S3/GCS), task creation, and result exports.Build a human-in-the-loop QC pipeline combining auto-validation and reviewer workflows.Create and maintain ML backend services for AI-assisted labeling using existing or custom models.2. Build Products that Convert Customer AI Needs into SolutionsWork directly with customers (internal/external) to translate AI training or labeling needs into product workflows using the annotation platform.Define ontology templates, QC workflows, and model integration plans for each customer.Prototype AI-assisted workflows quickly and iterate based on feedback.Manage the end-to-end lifecycle: requirements → design → implementation → validation → delivery.3. Integrate AI and MLOps StackConnect annotation outputs with downstream model pipelines (training, fine-tuning, retraining).Integrate evaluation metrics and feedback loops into product dashboards.Automate end-to-end flow from data labeling → model training → validation → retraining.Optimize compute, cost, and throughput across AI workloads.Required Skills and ExperienceExperience4–8 years of experience in full-stack or AI application development, with 2+ years in ML or AI-poweredsystemsProgrammingStrong in Python , Node.js , and cloud-native development (FastAPI, Flask, Express)AI ToolsExperience integrating vision, audio, and NLP models (AWS Rekognition, OpenAI Whisper, Pyannote, Hugging Face Transformers, Vertex AI, etc.)Data Annotation / QC ToolsHands-on with Label Studio , CVAT , Diffgram , or similar OSS platforms; experience extending them via API/SDKCloud PlatformsProficient in AWS or GCP (S3/GCS, Lambda/Cloud Functions, API Gateway, Cloud Run)Front-endFamiliarity with React / Next.js / Retool / Streamlit for internal dashboardsDatabasesPostgreSQL, DynamoDB, or Firestore; experience designing schema for annotation/QC metadataDevOpsDocker, CI/CD, environment management, API versioningAI Systems ThinkingAbility to design scalable, multimodal pipelines combining automation + human reviewSoft SkillsOwnership, clarity in communication, user empathy, ability to lead cross-functional effortsNice-to-Have SkillsKnowledge of MLOps orchestration tools (Airflow, Kubeflow, Prefect).Experience with active learning / data curation loops.Background in AI model evaluation or data quality metrics.Exposure to Reinforcement Learning from Human Feedback (RLHF) or LLM fine-tuning datasets.Prior experience building or managing an internal AI labeling workforce or QC automation system.Who You AreYou’re an engineer who loves building full systems , not just models — from UI to cloud to model integration.You have a strong product sense — you think in terms of user experience, feedback loops, and measurable outcomes.You’re hands-on and autonomous — able to ship production-quality code while owning design and strategy.You thrive in a startup-style environment where you define structure, not follow it.What we offerOpportunity to work at the cutting‑edge of GenAI training and evaluation for multimodal AI systems.Opportunity to directly work with global AI clients.Competitive salary + performance incentives.Fast tracked career growth in a rapidly expanding business within KGeN.For any questions, please write us at