01 About the role
We are hiring an AI/ML engineer to design and deploy production ML systems — not notebooks, not demos. You will build RAG pipelines, conversational AI agents, vector search systems, and the inference infrastructure that serves them to real users at scale.
This role sits in our AI Lab, where we build both internal tooling and client-facing AI products (AutoBot AI, RAG pipelines for support teams, custom LLM workflows). You will work directly with senior engineers and ship to production every week.
We care about real impact — models that actually reduce support tickets, retrieval systems that actually return the right document, agents that actually complete a task. If you have shipped an LLM feature to real users, we want to talk to you.
02 What we're looking for
- ✓ Strong Python (3.10+) — you write clean, typed, testable code, not just notebooks
- ✓ Hands-on experience with PyTorch or TensorFlow for model training and fine-tuning
- ✓ Production experience integrating LLMs — OpenAI, Anthropic, or open-source models via vLLM / Ollama
- ✓ Built and shipped RAG pipelines — chunking strategies, embeddings, vector stores (Pinecone, Weaviate, pgvector)
- ✓ LangChain or LlamaIndex for agent orchestration, tool use, and multi-step reasoning
- ✓ Prompt engineering as a real discipline — structured outputs, function calling, eval pipelines
- ✓ Vector databases and retrieval evaluation — recall, precision, reranking
- ✓ Familiarity with FastAPI or Flask for serving models
- ✓ Experience with model evaluation, A/B testing, and monitoring in production
- ✓ You read papers and can implement them — not just pip install someone else's wrapper
03 Tech stack you'll use
04 Why Softcracy
We're an engineering-first company — no PMs overriding engineers, no JIRA ticket mills. You'll own features end-to-end, ship to production weekly, and work alongside senior engineers who actually care about clean code. Remote-first, top-tier pay, and a real learning budget.
- ✓ Remote-first — work from anywhere in India
- ✓ Salaries above market median, transparent bands
- ✓ ₹25K/yr learning budget + conference passes
- ✓ Real ownership — you architect, build, and ship
- ✓ Stock options for every full-timer