Applied AI engineer
SiftHub · Mumbai, Maharashtra, India ·
- Category
- AI engineering
- Experience
- 3+ years
SiftHub · Mumbai, Maharashtra, India ·
Location: Mumbai,Maharashtra,India
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.
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.
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
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
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
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