AI / ML Engineer

PENN ENGINEERING; Bengaluru; On-site; INR 1,800,000 - 2,400,000; Full time

Job summary

PENN ENGINEERING is seeking an AI/ML Engineer to design and build AI foundations for its CX products, enabling customers to find fasteners and installation machines through natural conversations.

You will work on multimodal data ingestion, embeddings, vector stores, and retrieval-augmented decision making, deploying on AWS and collaborating with software and data teams to ship production-grade ML features.

Qualifications

• 5+ years of software engineering experience with 2+ years in production AI/ML or LLM-based systems.
• Strong embeddings expertise for text and multimodal data used in semantic search and recommendation.
• Experience with vector databases and retrieval-augmented patterns.
• Proficiency in Python and shipping ML/LLM features with testing and monitoring.

Responsibilities

• Ingest and structure product data from PIM sources for models.
• Understand product catalogs including drawings and images using vision/document models.
• Design embedding and retrieval systems and maintain vector stores.
• Build agentic experiences with tool use, planning, guardrails; production deployment on AWS.

Skills

Software engineering

Production ML

LLM

Embeddings

Vector databases

Retrieval patterns

Python

MLOps

AWS

Collaboration

Tools

Bedrock

SageMaker

OpenSearch

Lambda

ECS/EKS

Job description

As PennEngineering accelerates its Speed of Now transformation, the Customer Experience (CX) team is responsible for customer-facing digital products. We are looking for an AI / ML Engineer to help design and build the AI foundation and agentic capabilities that let customers find the right fasteners and installation machines, and receive a quote, through natural, conversational experiences. Behind those experiences sits a real machine learning and data challenge: ingesting and structuring product data from our PIM, interpreting the engineering drawings and product images in our catalogues, and matching all of it against a customer's stated requirement to recommend the correct part. This is a hands-on role that combines applied ML, multimodal and retrieval techniques, modern LLM and agent engineering, and disciplined production deployment on AWS. You will partner closely with the Software Architect, the Senior Data Architect, and the engineers building the CX product portfolio.

Key Responsibility

• Product Data & Multimodal Understanding
  • Build pipelines that ingest, clean, and structure product and attribute data from the PIM and other enterprise sources into a form that recommendation and retrieval models can use
  • Extract meaning from product catalogues that are not purely textual, including engineering drawings, dimensioned diagrams, and product images, using vision and document-understanding models
  • Design and maintain the embedding strategy for products, attributes, and customer requirements, including the choice of text and multimodal embedding models and how they are versioned and refreshed
  • Build and operate the vector stores and retrieval indexes that make product knowledge searchable by meaning, not just keywords
• Agentic Find, Quote & Recommendation
  • Build the agentic experiences that let a customer describe a need in their own words and be guided to the right fastener or installation machine, including tool and function calling, retrieval, multi-step planning, and guardrails
  • Develop the recommendation and ranking logic that matches a customer requirement against the product catalogue, combining semantic similarity, structured-attribute filtering, and compatibility rules
  • Build the agentic quoting flow on top of the recommendation layer so that a validated selection can move to a quote with minimal friction
  • Implement retrieval-augmented generation patterns that ground agent responses in accurate, current PennEngineering product knowledge
• AI Factory, Production & Integration
  • Develop and operationalise the AI Factory: reusable patterns for the feature store, model registry, prompt management, fine-tuning, evaluation harness, and monitoring
  • Implement responsible AI controls across all of the above: evaluation, red-teaming, content safety, observability, and cost guardrails
  • Partner with the Software Architect to integrate AI capabilities cleanly into the front end and microservices backend, and with the Senior Data Architect on the features, datasets, and embeddings these capabilities depend on
  • Ship ML and LLM features to production with proper testing and monitoring, and stay current with AWS's AI roadmap (Bedrock, SageMaker, AgentCore-style services) to recommend what to adopt and when

Requirements:

• Five or more years of software engineering experience, including two or more years building production AI/ML or LLM-based systems
• Strong, hands-on understanding of embeddings and embedding models, including how to choose, evaluate, and apply text and multimodal embeddings for semantic search and recommendation
• Experience with vector databases and modern retrieval patterns, including retrieval-augmented generationExperience building recommendation, matching, or ranking systems that combine semantic similarity with structured-attribute or rule-based filtering
• Working knowledge of computer vision or document-understanding techniques for extracting information from images, diagrams, or scanned technical documents
• Practical experience turning messy structured and semi-structured data (such as product catalogues or PIM exports) into clean, model-ready datasets
• Demonstrated experience building agentic systems: tool use, multi-step planning, evaluation, and guardrails
• Hands-on experience with AWS AI/ML services (Amazon Bedrock, SageMaker, OpenSearch or other vector stores, Lambda, ECS/EKS for inference)
• Strong Python skills and a track record of shipping ML/LLM features to production with proper testing and monitoring
• Working knowledge of feature stores, model registries, and MLOps tooling
• Comfortable collaborating across architecture, data, backend, and front-end teams

Preferred Qualifications:

• Experience with multimodal models that jointly reason over text and images, or with fine-tuning vision and embedding models for a specialised domain
• Familiarity with product-catalogue, configurator, fitment, quoting, or guided-selling use cases
• Experience operationalising LLMs in regulated, B2B, or enterprise environments
• Familiarity with manufacturing or industrial product data, including engineering drawings and dimensioned specifications
• AWS Certified Machine Learning Specialty or AI Practitioner certification
• Contributions to open-source AI/ML projects or applied research
• Recommendation quality: how accurately the system matches customer requirements to the correct fastener or installation machine
• AI feature quality: accuracy and evaluation benchmark scores for the agentic find and quote experiences
• Data and retrieval coverage: how completely product data and catalogue imagery are ingested, embedded, and made searchable
• Reliability in production: error rates, latency, and graceful failure handling of AI features
• Cost efficiency: inference and embedding cost per request, tracked and optimised against budget
• Speed to production: time to take a new AI capability from prototype to production-ready