AI/ML Developer

Horizon17 is hiring an AI/ML developer with 2–3 years of experience to build production AI modules for a document-intelligence and report-generation product.

About Horizon17

Horizon17 is a technology company. Our engineering teams build solutions across AI and machine learning, generative AI, blockchain, mobile, web, cloud and data.

We work in two ways. We build for clients, from strategy and design through engineering, launch and support. We also build and run our own in-house products, so we treat every system as owners would.

This role sits on one of those products. It reads unstructured documents and generates accurate, source-backed reports from them, with a human approving before anything is final.

What you will do

• Report generation: design and build LLM-powered generation of long-form, structured reports: sections, tables, citations and multi-format output.
• RAG pipelines: own document parsing, chunking, embeddings, retrieval, reranking and source citation, so every claim can show where it came from.
• Agent workflows: build planning, generation, regeneration and review flows in LangGraph and LangChain.
• Document intelligence: turn PDFs, scans, spreadsheets and other unstructured files into usable text and structure.
• Classification and scoring: build NLP and ML modules that classify, extract and score data points, with explainable outputs.
• Human in the loop: add conversational edits, clarification, approval, versioning and audit history, so missing or conflicting data triggers a question instead of an invented number.
• Production APIs: write Python services other product teams consume, including async jobs and streaming responses.
• Evaluation and quality gates: measure faithfulness and retrieval quality, run regression tests in CI, and monitor for hallucination and drift.
• Module ownership: take modules from design through production, write design docs, and take part in code reviews.
• Domain translation: work with product and domain experts to turn reporting rules into configuration and tests, so new document types do not need a rewrite.What you bring
• 2-3 years of professional engineering experience in Python (3.10+), with at least 1.5 years building LLM features that run in production.
• Production LLM work with OpenAI, Azure OpenAI, Anthropic or similar: prompting, structured output, retries, rate limits, and cost and latency trade-offs.
• Cost and token optimisation: has measured and cut LLM spend in production through prompt and context trimming, prompt caching, model routing (small model first, large model only when needed), batching, and per-feature token budgets with cost tracking.
• Hands-on RAG experience: embeddings, vector search, chunking strategy, reranking and retrieval evaluation.
• LangChain and LangGraph, or another agent orchestration framework, used for multi-step stateful workflows.
• Production backend skills: FastAPI or Flask, REST APIs, async Python, background jobs, and streaming over WebSockets or token streaming.
• Document AI and OCR: Azure Document Intelligence, Docling, PyMuPDF or similar.
• Measured LLM and RAG quality with RAGAS, DeepEval or equivalent, and pytest regression suites that gate releases.
• Strong data-layer skills: Pydantic, PostgreSQL, MongoDB and Redis.
• Has built and shipped at least one AI module end to end, from design to production.
• Fine-tuning or serving open-source models.
• Clear written communication: prompts, evaluation notes and technical design docs.Good to have
• ESG, sustainability or carbon accounting: GHG Protocol, Scope 1/2/3, BRSR, CDP, GRI or CSRD.
• High-stakes document products in finance, legal, healthcare or compliance reporting.
• Knowledge graphs (Neo4j or similar), or hybrid RAG and graph retrieval.
• Approval, audit and versioning workflows in an AI product.
• Docker, CI/CD and cloud deployment and monitoring of AI services.
• Classical ML where it fits: scikit-learn, XGBoost, sentence-transformers and explainability with SHAP.Role details

Item

Detail

Title

AI/ML Developer

Experience

2–3 years

Location

Gurgaon

Joining

Immediate joiners only

Product

Document-intelligence and report-generation product

Works with

AI/ML engineers, full-stack developers, DevOps, domain experts

LLM and retrieval

Python, OpenAI / equivalent LLMs, LangChain, LangGraph, embeddings and vector search

APIs and data

FastAPI / Flask, Pydantic, PostgreSQL, MongoDB, Redis

Documents and testing

Document parsing and OCR, pytest

Employment type

Full-time