AI Engineer (FoodTech Co.) (MUMBAI)
LabelBlind®️
Location
🇮🇳 Mumbai, India
Type
full_time
Salary
Undisclosed
Posted
2w ago
Job Description
AI Engineer – RAG & LLM Systems Location: Mumbai, In Office / Hybrid Experience: 3–5 Years
About the Role
: Company Description LabelBlind® Digital Solutions revolutionizes the food industry by offering comprehensive tools for Product and Labelling Compliance, Nutrition Assessment, Labelling Automation, and Market Readiness for food companies. Through its flagship product, FoLSol®️, LabelBlind® delivers India’s first digital food labelling solution designed to ensure regulatory compliance, transparency, accuracy, and efficiency in creating food labels. We are looking for a hands-on AI Engineer with strong expertise in the LangChain ecosystem to design, orchestrate, and optimize intelligent AI workflows.
Key Responsibilities
: - Design and build RAG pipelines for rule-based validation - Extract structured rules from PDF/XML/web sources using LLMs - Develop AI workflows using LangChain and LangGraph - Implement semantic search and embeddings for accurate retrieval - Use LangSmith for debugging, tracing, and evaluation - Prototype workflows using LangFlow - Generate explainable AI outputs for artwork validation - Optimize prompts and reduce hallucinations
Required Skills
: - Strong experience with LLMs and RAG systems - Hands-on expertise in LangChain, LangGraph, LangSmith, LangFlow - Experience with embeddings and vector databases (FAISS, Pinecone, Weaviate) - Proficiency in Python and NLP pipelines - Experience with unstructured data (PDF, HTML, XML) - Strong understanding of prompt engineering and evaluation
Good to Have
: - Experience in compliance or regulatory domain (Food labeling preferred) - Exposure to OCR tools (Tesseract, AWS Textract, Google Vision) - Knowledge of fine-tuning or domain-specific embeddings Success Metrics: - High accuracy in rule extraction and validation - Improved retrieval precision and reduced hallucinations - Clear traceability (rule → source → reasoning)