AI Intern in Hyderabad, भारत

About the role 

We're looking for an AI Intern to work alongside our engineering team on production machine learning systems, not side projects or sandbox demos. You'll ship code that real users touch: training and fine-tuning models, wrapping them in APIs, and helping take them from notebook to deployment. 

This is a 4-month structured internship with a clear path to a full-time offer. Interns who meet the performance bar at the end of the program are converted to full-time AI/ML Engineer roles. 

What you'll do 

• Build and maintain Python services and REST APIs using FastAPI to serve ML models in production 

• Work on computer vision problems, image classification, object detection, segmentation, using PyTorch/TensorFlow and OpenCV 

• Work on NLP tasks such as text classification, named entity recognition, summarization, and semantic search 

• Prototype and evaluate LLM-powered features: prompt engineering, RAG pipelines, embeddings and vector search, function calling, and fine-tuning where appropriate 

• Preprocess, clean, and annotate datasets; build reproducible data pipelines 

• Run experiments, track results, and clearly document what worked and what didn't 

• Benchmark model latency, cost, and accuracy, and help optimize on all three 

• Collaborate with engineering and product on scoping, code reviews, and deployment 

What we're looking for 

Required 

• Strong Python fundamentals, clean, readable, well-structured code 

• Hands-on experience with FastAPI (or Flask/Django with a willingness to pick up FastAPI quickly) 

• Practical exposure to computer vision and NLP, through coursework, personal projects, research, or a prior internship 

• Working knowledge of at least one deep learning framework — PyTorch or TensorFlow 

• Familiarity with LLMs and the modern tooling around them (OpenAI/Anthropic/open-source models, LangChain or LlamaIndex, vector databases) 

• Comfort with Git, and with reading and debugging code you didn't write 

• Solid grasp of ML basics: train/test splits, overfitting, evaluation metrics, and when a model is actually working 

• Currently pursuing or recently completed a Degree or Master Degree in CS, Data Science, Engineering, Mathematics, or a related field — or equivalent self-taught experience with a portfolio to show for it 

Nice to have 

• Docker and basic cloud experience (AWS, GCP, or Azure) 

• Experience deploying a model to production, at any scale 

• Familiarity with Hugging Face Transformers 

• Exposure to MLOps tooling, MLflow, Weights & Biases, DVC 

• SQL and general data-handling comfort 

• Open-source contributions, Kaggle results, or published research