Forward Deployed Engineer - GenAI

Systems Limited; Lahore; On-site; PKR 2,500,000 - 5,500,000; Full time

Job summary

Systems Limited is seeking a seasoned GenAI/LLM software engineer in Lahore to build production-grade GenAI applications, including copilot/chat experiences and enterprise search.

You will design RAG pipelines, fine-tune models (LoRA/QLoRA), and integrate OpenAI, Azure, AWS, and open-source endpoints. You’ll collaborate with data scientists and architects, focusing on reliability and cost control.

Qualifications

• 4–8 yrs software engineering with 1–3 yrs GenAI/LLM app building
• Strong Python; experience with LangChain and LlamaIndex
• Familiar with vector stores (Pinecone/Weaviate/pgvector) and knowledge graphs (Neo4j)
• Understanding of LLM failure modes and mitigation strategies
• Experience with model fine-tuning (LoRA/QLoRA) and evaluation
• Hands-on with enterprise GenAI/agentic platforms (Azure AI Foundry, AWS Bedrock, Vertex AI) and open-source tools

Responsibilities

• Build GenAI applications with LLM-powered features and enterprise search
• Design and implement RAG pipelines including chunking, embeddings, and re-ranking
• Fine-tune and adapt models (LoRA/QLoRA) as needed
• Engineer production prompts and prompt/context management
• Integrate LLM APIs with auth, rate limits, and cost controls
• Instrument apps for evaluation with logging and quality scoring
• Optimize latency and token cost via caching and model routing
• Translate client requirements into concrete GenAI feature specs
• Communicate tradeoffs to non-technical stakeholders
• Collaborate with architects, data scientists, and QA
• Document architecture and prompt design for handoff

Skills

Python

LangChain

LlamaIndex

LLM application building

prompt engineering

Tools

Pinecone

Weaviate

pgvector

Neo4j

Azure AI Foundry

AWS Bedrock

Vertex AI

LangSmith

Weights & Biases

Job description

ABOUT:

Builds generative AI applications — LLM-powered features, RAG pipelines, and enterprise search that ship to production, not just a demo.

KEY RESPONSIBILITIES

• Build GenAI applications — LLM-powered features, copilot/chat experiences, enterprise search
• Design and implement RAG pipelines: chunking strategy, embedding selection, hybrid retrieval, re-ranking, GraphRAG where structured retrieval is needed
• Fine-tune and adapt models (LoRA/QLoRA) when prompt engineering and RAG aren't sufficient
• Engineer and version production prompts; build prompt/context management into the application layer
• Integrate LLM APIs (OpenAI, Anthropic, Azure OpenAI) and open-source model endpoints with auth, rate-limiting, and cost controls
• Instrument applications for evaluation — output logging, quality scoring, human-feedback loops
• Optimize latency and token cost through caching, batching, and model routing strategies
• Translate client business requirements into concrete GenAI feature specifications
• Communicate technical tradeoffs (cost, latency, accuracy) to non-technical product stakeholders
• Collaborate with the Agentic AI Architect and Data Scientists on shared components
• Document architecture and prompt design decisions for handoff and maintainability

REQUIREMENTS & SKILLS

• 4–8 yrs software engineering, with 1–3 yrs hands-on GenAI/LLM application building
• Strong Python; experience with LangChain, LlamaIndex, or equivalent orchestration frameworks
• Vector databases and embedding strategies (Pinecone, Weaviate, pgvector), plus knowledge-graph/graph-database tooling (Neo4j) where relevant
• Understands LLM failure modes (hallucination, context-window limits, cost blowup) and designs mitigations
• Experience with model fine-tuning techniques (LoRA/QLoRA) and evaluation harnesses
• Hands-on with enterprise GenAI/agentic platforms — Microsoft Azure AI Foundry, AWS Bedrock (incl. Strands Agents SDK), and Google Vertex AI; open-source frameworks (LangChain, LlamaIndex) a good-to-have where no platform is mandated
• API design and integration experience, including auth, rate limiting, and streaming responses
• Familiarity with prompt-versioning and LLMOps tooling (LangSmith, Weights & Biases, or similar)
• Clear technical writing — documents a RAG architecture for a non-technical stakeholder
• Comfortable working directly with client engineers during embedded delivery
• Collaborative — works with architects, data scientists, and QA without needing everything pre-specified