About the Role

We are seeking a Full Stack Builder - AI Specialist to design, build, and deploy production-grade agentic AI applications and intelligent workflow systems who combines strong full-stack software development with hands-on generative AI implementation, requirements discovery, and technical solution architecture.

In this role, you will interface directly with business stakeholders to understand complex operational workflows (particularly across BFSI, supply chain, and industrial manufacturing), evaluate feasibility, design scalable technical blueprints, and execute end-to-end agentic and RAG solutions.

Key Responsibilities

1. Requirements Gathering & Technical Discovery

- Stakeholder Discovery: Conduct technical discovery sessions with product owners and domain stakeholders to unpack workflow bottlenecks and document clear functional/non-functional requirements.
- AI Feasibility & Risk Assessment: Critically assess whether a problem requires GenAI/agents or deterministic software/traditional ML, identifying edge cases, latency boundaries, and compliance risks early.
- Scope & Metric Definition: Define explicit success criteria (e.g., target accuracy, latency budgets, faithfulness thresholds, cost per transaction) and translate them into engineering backlogs and evaluation benchmarks.

2. Solution Design & Architecture

- End-to-End System Design: Architect resilient AI systems, producing clear technical design documents (TDDs), sequence diagrams, data flow schemas, and API contracts.
- Hybrid AI Strategy: Design architectures that intelligently blend open-source/open-weights models (via vLLM/Ollama) with proprietary frontier APIs based on data sovereignty, residency, and cost constraints.
- Security & Guardrail Design: Architect data sanitization layers, Human-in-the-Loop (HITL) confirmation flows, and role-based access controls (RBAC) to ensure sensitive business data remains protected.





3. Development & Execution

- AI-Native Development: Leverage AI engineering assistants (Claude Code, Cursor, Antigravity) to deliver robust, maintainable full-stack software quickly.
- Agentic Systems & Tool Calling: Implement multi-step agent workflows using open-source orchestration frameworks (LangGraph, LlamaIndex, or n8n) with strict structured outputs (Pydantic).
- Advanced RAG Pipelines: Build high-precision retrieval engines using vector databases (Qdrant, pgvector) with hybrid dense-sparse search, contextual re-ranking, and metadata filtering.
- Full Stack Integration: Build scalable backend services (Python/FastAPI, TypeScript, or C#/.NET) paired with reactive user interfaces (React/Next.js) supporting real-time streaming.
- Automated Evaluation: Set up automated CI/CD eval suites using frameworks like Ragas or DeepEval to catch hallucinations and regressions before production deployment.

Requirements & Qualifications

- Experience: 2+ years in software engineering with demonstrable experience building full-stack applications and deploying AI/LLM components.
- System Design & Architecture: Proven ability to translate business requirements into system architecture diagrams, database schemas, and API specifications.
- Core Stack: Strong proficiency in Python, TypeScript, or C#/.NET, along with relational databases (PostgreSQL/SQL Server) and open-source vector databases (Qdrant, pgvector).
- Agent & RAG Orchestration: Hands-on experience with LLM tool-calling, prompt engineering, structured output validation, and retrieval pipelines.
- Communication: Strong verbal and written communication skills to bridge technical architecture and business outcomes for non-technical stakeholders.

Nice-to-Haves

- Experience in data compliance/residency, Knowledge Graphs (Neo4j), or Industrial/Financial data schemas.
- Experience with on-premises/edge model deployment (Ollama, vLLM, LiteLLM).

📌 AI Full stack Developer (Mumbai)
🏢 Round Hole Square Peg (RHSP Solutions
📍 Mumbai