Jai Kisan
Website:
jai-kisan.com
Job details:
What You Will Do
API Design and Development: Utilizing Node.js for high-concurrency, real-time serving layers and Python for heavy data parsing, machine learning logic, and agent reasoning loops.
Agent Design: Architecting "Sense-Think-Act" loops utilizing advanced reasoning frameworks like ReAct, Plan-and-Solve, and Reflection to let agents independently navigate multi-step objectives.
Deterministic Tool & Function Calling: Engineering reliable agentic tool usage by mapping LLMs to Python/Node.js functions, ensuring precise parameter parsing and strict error-handling guardrails.
Stateful Multi-Agent Orchestration: Designing distributed, multi-agent systems with LangGraph or CrewAI to delegate specialized tasks (e.g., researcher, writer, validator) across collaborating AI personas.
Memory & Context Management: Implementing short-term (scratchpad/bufer) and long-term (vector database/episodic) memory structures to maintain coherent, deep-context agent interactions over extended sessions.
Production-Grade AI Evaluation: Proficient in setting up continuous evaluation pipelines measuring faithfulness, answer relevance, and factual correctness to ensure LLM outputs meet strict KPIs before deployment.
Robust Test-Driven AI: Combining traditional unit testing for API endpoints and custom Python tools with modern LLM evaluation frameworks to prevent regression in agentic behavior.
Financial Precision: Deep understanding of testing "Human-in-the-Loop" workflows, complex reconciliation logic, and "exception-clearing" AI where accuracy is non-negotiable.
Core Technical Skills
- Back-End Development: Advanced Node.js, Python, Fast API.
- AI Agent Orchestration: LangChain, LangGraph, Google ADK and smolagents.
- API & Integration: RESTful API design, OAuth, and External inference APIs.
- AI Engineering: Prompt engineering, RAG (Retrieval-Augmented Generation), function calling, and agentic tool usage.
- Testing & AI Evaluation: Unit testing (Jest/PyTest), agent evaluation frameworks (DeepEval, Ragas, TruLens), and LLM guardrails.
- Data & DevOps: SQL/NoSQL databases, Docker, uv package manager, and distributed tracing (Jaeger/Tempo).
- Experience: 3-6 years
Why Join CredServ?
You will be defining how global supply chains and enterprises evaluate and trust Artificial
Intelligence. If you want to work at the bleeding edge of Next-Gen Automation, Voice AI, and B2B
trade, this is your place.
- Competitive Salary & Equity
- Comprehensive Health & Wellness Benefits
Click on Apply to know more.