Qualifications
/
Full time
RAG and Retrieval Engineer
About Us
Whizzbridge is hiring a Mid to Senior RAG and Retrieval Engineer in Lahore. WhizzBridge is a technology solutions provider and a talent enabler. On one hand, it offers clients access to best of breed engineering talent and delivers their mission critical projects using the industry's best practices. On the other hand, it attracts and trains engineering talent on cutting edge technologies, programming languages and project management practices that set them up for successful professional and financial growth.
What We Offer
- Paid Leaves
- Medical Insurance
- Paid Udemy Courses and Certifications
- Career Progression Program
Job Description
- Design, build and tune retrieval augmented generation systems that ground language models in client proprietary data.
- Own the full retrieval pipeline end to end: ingestion, parsing, chunking, embedding, indexing, retrieval, reranking and generation.
- Select and evaluate embedding models against the client's actual corpus rather than against public benchmarks.
- Design chunking strategies appropriate to document type, including structured documents, long form text, tables and mixed media.
- Implement hybrid search combining vector similarity with keyword and metadata filtering.
- Implement reranking layers and measure their contribution to answer quality.
- Build and maintain retrieval evaluation suites measuring recall at k, context precision, groundedness and answer relevance, and run retrieval only evaluation before any generation evaluation.
- Detect, measure and reduce hallucination, including out of scope refusal and escalation routing.
- Plan and execute embedding model migrations and full corpus re indexing without service interruption.
- Optimise retrieval cost and latency at scale, including index selection, caching and query routing.
- Work directly with client subject matter experts to define what a correct answer looks like in their domain
Requirements
- Bachelor's degree in Computer Science, Software Engineering or a related field, or equivalent demonstrable experience.
- Three or more years of professional engineering experience, including hands on production RAG work. Tutorial level RAG is explicitly not sufficient for this role.
- Strong Python.
- Production experience with at least one vector database, such as pgvector, Pinecone, Qdrant, Weaviate or Chroma, including indexing strategy and query tuning.
- Demonstrated understanding of chunking strategy, embedding model selection and the trade offs between them.
- Practical experience with hybrid search and reranking.
- Experience with a retrieval orchestration framework such as LlamaIndex, LangChain, LangGraph or DSPy.
- Experience running structured evaluation of a retrieval system, with numbers attached.
- Strong written and verbal English.
Qualifications
- Experience with Ragas, LangSmith, DeepEval or a comparable evaluation framework.
- Experience with graph databases such as Neo4j or Memgraph for knowledge graph augmented retrieval.
- Experience with document parsing at scale, including OCR, PDF extraction and table extraction.
- Experience with multimodal retrieval across text, image or audio.
- Familiarity with data privacy and access control in retrieval, including row level and document level permissions.
- Experience in a regulated domain such as legal, healthcare or financial services.
- Exposure to agentic retrieval patterns and query planning.