AI Engineer

ALP (Computational Finance, AI and Data Engineering); Bengaluru; On-site; INR 1,200,000 - 1,800,000; Full time

Job summary

Angel Lane Partners (ALP) is building end-to-end LLM applications for banking and regulatory clients from Bengaluru with frequent GCC travel. You will develop RAG pipelines, LangChain workflows and deploy LLMs for on-premise/private-cloud environments, while collaborating with data engineers and domain experts.

The role involves extracting data from financial documents, evaluating model outputs and ensuring security with guardrails and access controls, with travel to client sites as needed.

Qualifications

• 2–4 years in software or ML engineering with at least 1 year building LLM applications.
• Strong Python, including API development and asynchronous code.
• Experience with LangChain, RAG, and vector stores.
• Hands-on with ChromaDB, Pinecone and pgvector in PostgreSQL.
• Familiarity with hosted LLM APIs and open-weight small models.
• Ability to build evaluation sets and measure LLM output quality.
• Docker, Git and REST API experience.

Responsibilities

• Build RAG pipelines covering ingestion, parsing, chunking, embedding and retrieval.
• Develop LLM workflows in LangChain with prompts, tools and multi-step agents.
• Set up and manage vector stores with configuration and filtering.
• Deploy LLMs for on-premise/private-cloud via vLLM or similar.
• Extract structured data from PDFs and documents, validating against sources.
• Create evaluation sets and track retrieval quality and citation accuracy.
• Implement guardrails: filtering, grounding checks, PII masking and access control.
• Expose LLM features via APIs for client applications.
• Log prompts, responses, latency and costs; monitor production behavior.
• Store metadata and audit records in PostgreSQL; document model choices.

Skills

Python programming

API development

LangChain

RAG architecture

ChromaDB

Pinecone

PostgreSQL with pgvector

SQL and database design

LLM applications in production

Docker, Git, REST APIs

Hosted LLM APIs & open-weight models

Education

Bachelor's or master's degree in Computer Science / AI / Data Science

Tools

LangChain

RAG pipelines

ChromaDB

Pinecone

PGVector

FastAPI

Docker

Git

REST APIs

Job description

Location: Bengaluru, with frequent GCC travel

About ALP

Angel Lane Partners (ALP) is a computational finance, data engineering and AI/ML firm with offices in Dubai, London, Bengaluru and Riyadh. We work with banks, financial institutions and government entities across the GCC.

Role Overview

You will build LLM applications for ALP's banking and regulatory clients, such as extracting data from credit files and financial statements, answering questions on internal policies and regulations, and drafting reports. Many clients require private or on-premise deployment, so you will work with small open-weight models as well as hosted APIs. You will work with data engineers, full-stack developers and domain specialists, and support deployments in client environments.

Travel

The role is based in Bengaluru, with frequent travel to client sites in the GCC, sometimes for several weeks at a time. A valid passport is required.

Key Responsibilities

• Build RAG pipelines covering document ingestion, parsing, chunking, embedding, retrieval and re- ranking.
• Develop LLM workflows in LangChain, including prompt templates, tool calling, structured outputs and multi-step agents.
• Set up and manage vector stores (ChromaDB, Pinecone, PostgreSQL with pgvector), including index configuration, metadata filtering and hybrid search, and choose the store that fits each deployment.
• Deploy small language models for on-premise and private-cloud use, including model selection, quantisation and serving through vLLM, Ollama or similar.
• Extract structured data from PDFs, scanned documents and spreadsheets, such as financial statements, credit memos and regulatory filings, and validate outputs against source.
• Build evaluation sets with domain specialists; measure retrieval quality, answer accuracy, citation accuracy and hallucination rates, and track them across releases.
• Implement guardrails: input and output filtering, grounding checks, PII masking and access control on retrieved documents.
• Expose LLM features through APIs that full-stack developers integrate into client applications.
• Log prompts, responses, latency and token cost; monitor production behaviour and investigate failures.
• Store application data, document metadata and audit records in PostgreSQL.
• Document model choices, prompts, evaluation results and known limitations for client review.

Required Skills

• 2–4 years in software or ML engineering, with at least 1 year building LLM applications in production or client pilots.
• Strong Python, including API development (FastAPI or Flask) and asynchronous code.
• LangChain and RAG architecture: chunking strategies, embedding models, retrieval and re-ranking.
• Hands-on ChromaDB, Pinecone and PostgreSQL with pgvector.
• SQL and PostgreSQL schema design.
• Experience with both hosted LLM APIs and open-weight small language models.
• Building evaluation sets and measuring LLM output quality.
• Docker, Git and REST APIs.

Preferred Skills

• Fine-tuning with LoRA or QLoRA.
• OCR and document layout parsing for scanned files and tables.
• RAG evaluation frameworks such as RAGAS.
• LangGraph or LlamaIndex.
• Certifications in BI, ML, cloud or infrastructure.
• GPU serving and inference optimisation.
• Banking or financial-services documents: financial statements, credit files, regulatory circulars.

Qualifications

• Bachelor's or master's degree in Computer Science, AI/ML, Data Science or a related field.
• At least one LLM application delivered beyond prototype stage.
• Able to explain model behaviour and limitations to non-technical client stakeholders.