Lead AI Engineer

Company Description

The Ksquare Group is a consulting company focused on delivering cutting-edge technological solutions through customized projects that enhance decision-making with speed and quality. By partnering with leading technology platforms, the company accelerates digital transformation initiatives for its clients. Its core expertise includes Salesforce and data solutions, helping organizations modernize processes and leverage insights for better business outcomes. Professionals joining The Ksquare Group can expect to work in a dynamic environment centered on innovation, collaboration, and high-impact technology delivery.

Role Summary

• Design and build the AI platform core model routing/gateway layer, prompt/version management, RAG pipelines, embedding infrastructure, and vector store operations at scale.
• Own model lifecycle infrastructure fine-tuning pipelines, evaluation harnesses, A/B testing frameworks for model/prompt changes, and regression detection when models or prompts change.
• Build serving infrastructure low-latency inference paths, caching layers, batching strategies, fallback/routing logic across multiple model providers (OpenAI, Anthropic, open-source/self-hosted).
• Establish observability for AI systems tracing, token/cost tracking, hallucination/quality monitoring, drift detection — the stuff that’s invisible until it breaks in prod.
• Create internal developer tooling SDKs, APIs, and abstractions so product teams can consume AI capabilities without reinventing infra each time.
• Data and retrieval infrastructure chunking strategies, embedding pipelines, hybrid search, freshness/index management for RAG systems.
• Guardrails and safety infra content filtering, PII redaction, rate limiting, abuse detection at the platform layer.
• Cost and performance optimization model selection strategy, quantization/distillation where relevant, caching, and spend visibility across teams.
• Partner directly with product/ML teams to understand real usage patterns and evolve the platform based on actual production pain, not theoretical needs.

What You Need to Bring (Non-Negotiable)

• Strong software engineering fundamentals — this is an infrastructure/platform role first, AI second. You should be comfortable owning distributed systems, APIs, and production reliability.
• Hands-on experience building (not just using) at least one of: RAG pipelines, model serving infra, fine-tuning pipelines, or LLM orchestration frameworks.
• Solid understanding of how LLMs actually work under the hood — tokenization, context windows, embeddings, attention basics — enough to make real tradeoffs, not just quote them.
• Experience with at least one vector database (pgvector, Pinecone, Weaviate, Qdrant, etc.) in a real production setting.
• Production experience with observability/monitoring for ML or AI systems.
• Comfortable in Python (and ideally one of Go/Rust/TypeScript for platform-side services).
• Experience with cloud infra (AWS/GCP/Azure) and containerized deployment (Docker/Kubernetes).