Website:
wayudhan.in
Job details:
Role Description
This is a full-time, on-site Senior AI Engineer role based in Gurugram. As Senior AI Engineer —— you will build the AI agent core. You will work directly with the CTO on architecture and with the CEO on product requirements, translating developer pain points from discovery calls into production agent workflows.
You will have real influence on how the platform is architected, which models we use, and how we think about agent reliability. That influence comes with early-employee equity and the opportunity to grow into a Staff or Principal Engineer role as the team scales.
Qualifications
Must-Have — Non-Negotiable
• 3+ years building production LLM applications: not demos, not side projects — systems with real users, real load, and real failure modes. You have debugged a hallucination in production at 2am and know exactly how to prevent it next time
• Strong Python engineering: FastAPI or Flask, async/await patterns, PostgreSQL, Redis. You write code that your future colleagues will thank you for, not curse you for
• Hands-on agent framework experience: LangChain, LlamaIndex, AutoGen, or similar. You have built agents that use tools, maintain memory, and hand off between steps — not just called the OpenAI chat completion endpoint
• API integration experience: you have built and maintained production API integrations that handle auth, retries, rate limits, and edge cases. The WhatsApp Business API will test all of these
• Self-directed and ownership-minded: at a 3-person company, nobody will assign you tickets. You will read discovery call notes, identify the biggest agent reliability gap, and fix it before being asked.
Strong Advantage — Will Prioritise
• WhatsApp Business API in production: you have built a WhatsApp bot that processes thousands of messages per day. You know about template message approval, webhook verification, media handling, and the frustrating 24-hour session window
• Multilingual / Indic NLP experience: you have worked with Hindi, Marathi, Gujarati, or other Indic languages in an NLP context. You know the difference between transliteration and translation and why it matters for intent classification
• LLM fine-tuning experience: LoRA, QLoRA, or full fine-tuning on domain-specific corpora. You know how to construct a training dataset, evaluate BLEU/ROUGE, and avoid catastrophic forgetting
• Vector databases and RAG: Pinecone, Weaviate, or Chroma in production. You have built a RAG pipeline that actually retrieves the right chunks and know the difference between naive chunking and semantic chunking
• Government API / scraping experience: you have worked with Indian government APIs or built scrapers for government portals. You understand the quirks of public sector data: inconsistent formats, authentication surprises, and unpredictable downtime
Nice to Have
• Open-source contributions: a public GitHub with AI/NLP work is a strong positive signal. Not required, but it removes ambiguity about your experience level
• Experience with Sarvam AI or Krutrim: India-specific LLMs that are increasingly relevant for vernacular agent deployments
• Cloud infrastructure comfort: you can deploy a Docker container to AWS ECS, set up an RDS instance, and read a CloudWatch alarm. You do not need to be a DevOps expert — the CTO owns infra — but you should not be blocked by it
Click on Apply to know more.