AI/ML Engineer I

Job descriptionJob Role - AI/ML Engineer IJob Type: Full-Time Mode: WFO - Onsite

Location: PuneNotice Period: Immediate (preferred) to 30 daysExperience: 01 to 03 yrsMost engineers with a year of AI experience have trained models.

Few have kept one alive in production. We're hiring for the second kind.Plenty of places will let you fine-tune a model on a clean dataset, report an accuracy number and call it done. We'd rather hand you a real client problem, a senior engineer who will tell you honestly when your approach won't hold up, and enough rope to own a scoped piece of the system end to end.If you want the next year to be about production judgement rather than another notebook, this is the right place for it.About Kreeda LabsWe are an AI and full stack product engineering company based in Pune. We have delivered 75+ production systems for 40+ clients across three continents, spanning fashion, HR tech, healthcare, insurance, eLearning and supply chain.Our work includes engagements with Ralph Lauren, Tata Group / Tata Play, Patanjali and Multiversity. AI is not a side experiment here — it is what a growing share of our client work is built on, and this role works on it directly.What you'll doYou will work alongside senior engineers on live client systems, not sandboxed exercises — starting with well-scoped pieces and taking on more as you prove you can carry them.Build

• Build and train ML/DL models for real client use cases, from data preparation through evaluation
• Build Computer Vision pipelines, including image preprocessing and object detection
• Prototype GenAI and RAG applications, including vector database integrationsShip
• Build production inference pipelines and REST APIs (FastAPI) around your models
• Deploy and maintain AI applications on AWS using Docker
• Debug and stabilise deployed models under real traffic, not just in a notebookResearch and improve
• Fine-tune open-source models and compare approaches, including PEFT/LoRA
• Read papers, documentation and open-source repos, and turn promising ideas into working prototypes
• Evaluate models and systems properly, and use failure cases to improve themWho can applyMust have
• Approximately 1 year of hands-on engineering experience, including real AI/ML systems
• Strong Python, with scikit-learn and PyTorch and/or TensorFlow
• Practical experience building, training and evaluating ML/DL models, including data preprocessing and feature engineering
• Computer Vision fundamentals: OpenCV/PIL, image preprocessing and object detection (YOLO or similar)
• Hands-on GenAI work: prompt engineering, Hugging Face/Transformers, RAG and vector databases
• Fine-tuning experience with open-source models, including PEFT/LoRA familiarity
• Agent and tool-calling experience with LangChain, LangGraph or an equivalent framework
• Backend and API development with FastAPI, plus Pydantic/request validation
• Cloud deployment experience with Docker, AWS S3 and AWS EC2
• Ability to evaluate models with the right metrics and explain what a failure case taught you
• The ability to explain your own architecture and model choices without hiding behind jargonGood to have
• MLOps tooling: MLflow, CI/CD, Kubernetes
• AWS ECR / ECS / CloudWatch
• Multimodal or Vision-Language Models, advanced model optimisation or quantisationYou don't need every line here. What matters is that you can point to at least one AI/ML project you personally took into production — not just trained.What makes this differentAt most companies, a year into an AI role you are still working in one corner of one product, off someone else's spec. Here you will touch real client systems across industries within your first year — the fastest way to build the production judgement that separates an AI/ML engineer from someone who can only train models.You will also work in an engineering environment where AI-assisted development is normal. We expect you to use those tools well and still be able to review and defend everything that ends up in the repository.What you'll gain
• Hands-on ownership of scoped AI/ML components inside real client systems, not sandboxed exercises
• Exposure to production GenAI, Computer Vision and classical ML work across industries, not one product
• Direct mentorship from senior engineers who will tell you honestly when your approach won't hold up
• A clear, merit-based path toward Senior AI Engineer, defined by what you ship and fix
• Compensation that reflects hands-on production experience, not just a generic entry-level bandThis role may not be for you ifYou want to stay in notebooks and hand off anything that touches production, you are only comfortable with clean tutorial datasets, you expect to be told exactly what to build rather than figuring it out, or you have not yet shipped anything a real user depended on.We aim to close the process within two weeks and we respond either way.