AI Engineer (GenAI Applications)

ABOUT THE ROLEWe're seeking a skilled AI Engineer with 4+ years of experience to build applications powered by Generative AI — integrating foundation models via APIs and fine-tuning lightweight open-source models for specific use cases. You'll work with cross-functional teams to develop and deploy AI-powered features across our product.KEY RESPONSIBILITIESDesign and build AI-powered features using GenAI models (OpenAI, Anthropic, Gemini, etc.) via APIs, and fine-tune/adapt open-source models (LoRA/QLoRA, PEFT) for domain-specific needsBuild prompt engineering and RAG pipelines (vector databases like pgvector, Pinecone, Weaviate) to ground outputs in proprietary dataDesign and orchestrate multi-step AI workflows/agents (tool use, function calling, chaining)Own the end-to-end AI feature lifecycle — data prep, model/API selection, evaluation, integration, deployment — in collaboration with engineering and product teamsMonitor and optimize model performance, latency, and cost; ensure ethical AI practices (fairness, transparency, data privacy.REQUIRED QUALIFICATIONSBachelor's/Master's in Computer Science, AI/ML, or related field (or equivalent experience)4-8 years of hands-on experience building applications using GenAI/LLM APIs (OpenAI, Anthropic, Google, etc.)Strong Python skills, with experience in prompt engineering, structured outputs, function/tool calling, and orchestrating AI APIs in productionExperience fine-tuning open-source models (Hugging Face, LoRA/QLoRA/PEFT) and building RAG pipelines (embeddings, vector search, retrieval evaluation)Familiarity with agent frameworks (LangChain, LangGraph, LlamaIndex, or similar) and cloud platforms (AWS, GCP, Azure)PREFERRED SKILLSModel evaluation frameworks (LangSmith, Ragas) and MLOps/LLMOps practices (prompt versioning, drift monitoring, CI/CD for AI features)Computer vision APIs (OCR, image generation, vision-language models) for multimodal use casesKnowledge of responsible AI practices (bias mitigation, guardrails, hallucination detection) and self-hosted inference (vLLM, Ollama)Experience: 4-8 yearsLocation: Bodakdev, Ahmedabad (WFO)