India onlyOur PeopleWith over 1,500 team members across 15+ countries, we operate in a global, remote-first environment. We are building more than software; we are building a global community rooted in creativity, collaboration, and impact. We take pride in cultivating a culture where innovation thrives, ideas are celebrated, and people come first, no matter where they call home.
Our ImpactAs of mid 2025, our platform powers over 1.5 billion messages, helps generate over 200 million leads, and facilitates over 20 million conversations for the more than 2 million businesses we serve each month. Behind those numbers are real people growing their companies, connecting with customers, and making their mark - and we get to help make that happen.
About the RoleWe’re looking for a lead-level engineer to own and drive Gen AI initiatives for Workflows designing, building, and scaling AI-powered workflow automation features end-to-end. You’ll integrate and orchestrate LLMs (OpenAI, Anthropic, Claude, LLaMA, Mistral), design RAG pipelines with vector DBs (Pinecone, Weaviate, FAISS, Qdrant), and implement agentic AI workflows using LangChain, LangGraph, and AutoGen.You’ll lead the technical strategy for AI in workflows from prompt engineering and fine-tuning (LoRA, QLoRA, PEFT) to EVAL frameworks that ensure performance, accuracy, and reliability. Our stack: Vue, Node.js, MongoDB, Firestore, ElasticSearch, Redis, all on GCP and we expect you to move seamlessly across it while delivering scalable, production-grade AI capabilities.
Requirements:
- Leadership & Ownership: Proven experience leading projects and mentoring engineers
- Fullstack Engineering: Node.js backend, Vue frontend, API design & scaling
- LLM APIs: OpenAI, Anthropic, Bedrock, Claude, LLaMA, Mistral, LlamaIndex
- Prompt Engineering: few-shot, CoT, dynamic context, function/tool calling
- RAG Pipelines: embeddings, chunking, hybrid retrieval
- Vector DBs: Pinecone, Weaviate, FAISS, Qdrant
- Agentic AI: multi-step workflows, tool usage, memory (LangChain, LangGraph, AutoGen)
- Fine-Tuning: LoRA, QLoRA, PEFT, embedding-based adaptation
- EVALs: hallucination detection, groundedness, BERTScore, BLEU, GPTScore
- Data Layer: MongoDB, Firestore, ElasticSearch, Redis
- Cloud & Deployment: GCP (preferred), AWS, or Azure; Docker/Kubernetes; CI/CD for AI workloads
Responsibilities:
- Lead the architecture, development, and deployment of AI-powered workflow automation features
- Integrate and orchestrate multiple LLM providers with high reliability
- Build and optimize RAG pipelines embeddings, chunking, hybrid retrieval using vector databases
- Design and implement agentic AI workflows multi-step reasoning, tool usage, memory leveraging LangChain, LangGraph, AutoGen, or similar
- Apply prompt engineering best practices: few-shot, chain-of-thought, dynamic context, function/tool calling
- Execute fine-tuning/adaptation workflows LoRA, QLoRA, PEFT, and embedding-based customization
- Develop and manage EVAL frameworks hallucination detection, groundedness, BERTScore, BLEU, GPTScore to track and improve model quality
- Build robust, scalable Node.js APIs and Vue-based frontends, integrating seamlessly with our stack
- Work with MongoDB, Firestore, ElasticSearch, Redis to design efficient data flows for AI workloads
- Ensure observability, performance tuning, and cost optimization in GCP-based deployments
- Collaborate closely with product, design, and infra teams to deliver high-impact features on time
- Mentor engineers, set best practices, and drive technical decision-making in the AI workflows domain
#NJ1
United States
AF, AM + 49 more
