Website:
lemici.com
Job details:
About UsWe are an early-stage AI-first startup building intelligent solutions. Over 60% of our product is powered by AI and we are actively fine-tuning Small Language Models (SLMs) and Large Language Models (LLMs) for various real-world use cases, alongside building robust data pipelines to power them. You won't be fetching coffee here — you'll be working on things that actually ship.
Role OverviewWe are looking for a motivated and curious MLOps Intern to join our growing tech team. You will work closely with our AI engineers and DevOps team to help operationalize our LLM and SLM workflows — from data preparation and fine-tuning pipelines to model deployment and monitoring in production. This is a hands-on LLMOps role where you'll get exposure to the full lifecycle of language model systems.
If you're a student who is excited about LLMs, SLMs, and the infrastructure that makes AI products actually work in production — this is the role for you.
What You'll Work On· Help build and maintain data pipelines for curating, cleaning, and versioning datasets used for SLM/LLM fine-tuning
· Assist in setting up and maintaining experiment tracking for fine-tuning runs using tools like MLflow or Weights & Biases
· Support model versioning and registry workflows to manage multiple fine-tuned SLMs across different use cases
· Work with the DevOps team to containerize and deploy LLM inference workloads using Docker and Kubernetes
· Assist in setting up LLM observability and monitoring — tracking latency, output quality and model drift in production
· Help evaluate and integrate inference optimization techniques like quantization, GGUF or vLLM for efficient model serving
· Support RAG (Retrieval-Augmented Generation) pipelines — including vector database integrations and chunking strategies
· Write documentation for LLMOps processes and pipelines
· Collaborate with AI engineers and DevOps interns on day-to-day tasks
What We're Looking ForMust Have:
· Currently pursuing a B.Tech / B.E. / M.Tech / M.Sc in Computer Science, Data Science, AI/ML or a related field
· Basic understanding of how Large Language Models and fine-tuning work conceptually
· Familiarity with Python and the HuggingFace ecosystem (Transformers, Datasets, Tokenizers)
· Understanding of what a data pipeline is and how data flows through an AI system
· Willingness to learn and get hands dirty with new tools quickly
Good to Have:
· Hands-on experience fine-tuning or working with open-source LLMs/SLMs (Mistral, LlamA, Phi, Gemma, etc.)
· Familiarity with LangChain/LangGraph, LlamaIndex or similar LLM orchestration frameworks
· Basic knowledge of Docker (building and running containers)
· Exposure to tools like MLflow, Weights & Biases or DVC
· Understanding of RAG pipelines and vector databases (FAISS, Pinecone, ChromaDB etc.)
· Awareness of model inference tools like vLLM, Ollama or llama.cpp
· Familiarity with cloud platforms (AWS / GCP / Azure) even at a conceptual level
What You'll Gain· Real-world experience working on production LLM/SLM systems
· Deep hands-on exposure to fine-tuning Small Language Models
· Experience with the full LLMOps lifecycle — data, fine-tuning, deployment, monitoring
· Certificate of Internship for your resume and LinkedIn
· Potential for Pre-Placement Offer (PPO) based on performance, as we scale
Click on Apply to know more.