Job description
This role is for one of the Weekday's clients Salary range: Rs 5000000 - Rs 7000000 (ie INR 50 - 70 LPA) Min Experience: 5+ years Location: Noida, Uttar Pradesh, India JobType: full-time As a core engineer in a brand-new 0-to-1 vertical build , you will design, build, and ship production-grade, LLM-powered systems tailored for personalized learning in Indian languages and local educational contexts.
You will own applications end-to-end: RAG pipelines, agentic workflows, evaluation harnesses, and the production microservices that serve them.
We maintain a strict bar for robust software engineering , not just prompt crafting or notebook experimentation.
This role requires high ownership, comfort with ambiguity, and hands-on execution during our early vertical setup phase.
Key Responsibilities Production LLM Pipelines: Architect and deploy robust LLM applications (RAG architectures, multi-step agentic workflows, and tool-calling systems) for use cases like automated question answering, adaptive feedback, and curriculum alignment.
Full-Stack ML Systems: Own data pipelines, retrieval layers, orchestration layers, APIs, and service infrastructure—written as clean, maintainable, and thoroughly tested production code.
Rigorous Evaluation Frameworks: Build task-specific benchmarks, regression testing, human-in-the-loop evaluation loops, and automated quality gates to systematically prevent model degradation.
Latency & Cost Engineering: Implement prompt optimization, caching strategies, request batching, and model routing to ensure production systems remain fast and financially viable at a population scale.
Indic Language Grounding: Build and scale data preparation pipelines for instruction data, with a specific focus on multilingual and Indic language sources.
Technical Profile Required Must-Have Skills: Production-Grade Python: Strong background in writing clean, modular, and tested Python code.
Experience owning live services in production is non-negotiable (no notebook-only engineers).
RAG & Agent Orchestration: Deep experience building and shipping RAG or agentic systems using production-proven frameworks (e.
g., LangChain, LangGraph, or custom equivalents), with strong tool-calling and multi-step reasoning design.
Automated Evaluation: Track record of building scientific evaluation workflows, systematic regression testing, and quality monitoring for LLM outputs.
Backend Integration: Practical experience with API development, backend data pipelines, and cloud deployment infrastructure.
Good-to-Have Skills: Model Fine-Tuning: Hands-on experience with LoRA, QLoRA, SFT, or DPO preference optimization.
Inference Optimization: Experience reducing latency and serving costs via quantization (AWQ/GPTQ), ONNX compilation, or serving frameworks like vLLM/TGI.
Distributed Frameworks: Familiarity with DeepSpeed or FSDP for handling training datasets at scale.
What We Offer Population-Scale Impact: Directly shape how millions of children across India learn and thrive within public education systems.
Autonomy of a 0-to-1 Build: Enjoy the agility and growth potential of building a team and product from scratch, backed by the stability of a highly successful, profitable parent company.
Comprehensive Benefits: Competitive compensation package, supportive work culture, and employee-centric health insurance benefits.
Must-have skills Python, rag, llm Good-to-have skills peft, Quantization, onnx
This job post has been translated by AI and may contain minor differences or errors.
