Applied AI engineer

SiftHub is an Agentic Platform for Deal Orchestration. We deploy agents to automate sales collateral, deal briefs, RFPs, and post-meeting follow-ups; unblocking deals for sales and solutions teams.SiftHub collates and sifts through your ecosystem (Drive, Gong, CRM, Slack, and more) to identify and convert real-time deal signals into tailored content and ready-to-execute actions. Designed to replicate how top-performing reps execute deals. Trusted by revenue teams at Saviynt, Everlaw, Allego, Superhuman, Sirion and more.

Why This Role Exists SiftHub's core value is the quality of the intelligence it surfaces to sales teams - delivered through a web app, Chrome extension, Microsoft add-in, and Google Sheets integration. That intelligence comes from a combination of retrieval, reasoning, and generation and it only works if every layer is built and evaluated with care. We need an Applied AI engineer who can design and own these pipelines end to end: not just get them working, but get them reliably good and measurably improving over time.

What You'll Own

• End-to-end RAG pipelines: retrieval architecture, chunking strategy, re-ranking, and prompt design
• Evaluation frameworks: defining quality metrics, building eval harnesses, and tracking pipeline health over time
• Agentic workflows and LLM integration: multi-step reasoning, tool use, orchestration, model selection, context management, latency, and cost optimisation
• Collaboration with Backend engineers to serve NLP outputs at production quality and latency
• Research-to-production translation: staying current and knowing what’s worth shipping

Experience & Core Skills

• Bachelor’s or Master’s degree in Computer Science or a related field, or equivalent practical experience
• 3–6 years of production NLP/ML experience with shipped models or pipelines used by real users
• Hands-on experience building and improving production-grade RAG systems
• Experience defining quality metrics and building evaluation systems for LLM outputs
• Solid Python engineering fundamentals with production-grade code (not just notebooks)
• Experience with LangChain, LlamaIndex, FastAPI, or similar agentic frameworks
• Experience working with LLMs such as OpenAI GPT-4 or Anthropic Claude
• Familiarity with vector databases like Pinecone, ElasticSearch or pgvector
• Experience deploying or working with cloud platforms such as AWS or Azure

Good to Have

• Experience with multi-agent orchestration patterns
• Familiarity with enterprise data sources (heterogeneous formats, noisy inputs, schema variation)
• Exposure to latency and cost optimisation for LLM-powered products
• Background in sales tech, CRM data, or B2B SaaS environments

What We Offer

• Work on genuinely hard AI problems, not CRUD apps
• Direct access to founder
• Access to best-in-class dev tools and AI assistants to help you do your best work
• Health insurance
• In-office in Mumbai, we value building together in person