Job Position - Gen AI Developer
Company- RiDiK Pte Ltd
Experience - 5+ Years
Location - Dhahran, Saudi
Job Description -Strong foundation in deep-learning mathematics (linear algebra, probability, optimization). Expert in PyTorch and/or Tensor Flow. Mastery of Hugging Face Transformers/Diffusers and ability to write custom CUDA kernels. Experience with distributed training on GPU/TPU clusters. Proficiency with prompt-engineering tools (Lang Chain, Llama Index). RAG expertise: building vector stores, hybrid retrieval, and multi-agent orchestration. MCP server knowledge: designing, deploying, and scaling services that manage agent communication and state (REST/g RPC, message queues, load balancing). MLOps stack: Docker, Kubernetes, Helm, CI/CD, Triton/Tensor RT-Inference Server, Prometheus, Grafana. Large-scale data pipelines (Spark, Beam, Dask) and data versioning (Lake FS, DVC, Weights & Biases). Publications or patents in generative AI (Neur IPS, ICML, ICLR, CVPR). Open-source contributions to major libraries (Hugging Face, PyTorch, Diffusers). Domain expertise in finance, healthcare, gaming, or creative media. Roles & Responsibilities :Design, train, and fine-tune transformer, diffusion, and multimodal models of at least 1 billion parameters. Implement distributed training on GPU/TPU clusters using appropriate scaling frameworks. Build prompt-engineering pipelines (few-shot, chain-of-thought, retrieval-augmented generation). Utilize libraries such as Lang Chain, Lang Graph, Llama Index, or equivalents. Architect and implement RAG pipelines that combine vector stores, external knowledge bases, and multiple specialized agents (e.g., summarization, code generation, data extraction). Deploy and manage MCP (Multi-Channel/Component Processing) servers to coordinate agent communication, routing, and state management. Provide APIs and SDKs for downstream services to invoke specific agents or composite RAG workflows. Construct petabyte-scale multimodal data pipelines (Spark, Beam, Dask). Apply data-centric AI practices and versioning (Lake FS, DVC, Weights & Biases). Containerize services (Docker) and orchestrate with Kubernetes/Helm. Set up CI/CD pipelines (Git Hub Actions, Git Lab CI) and serve models via Triton, Tensor RT-Inference Server, or Torch Serve. Implement observability (Prometheus, Grafana, Open Telemetry).

About CLPS RiDiKRiDiK is a global technology solutions provider and a subsidiary of CLPS Incorporation (NASDAQ: CLPS), delivering cutting-edge end-to-end services across banking, wealth management, and e-commerce. With deep expertise in AI, cloud, big data, and blockchain, we support clients across Asia, North America, and the Middle East in driving digital transformation and achieving sustainable growth. Operating from regional hubs in 10 countries and backed by a global delivery network, we combine local insight with technical excellence to deliver real, measurable impact. Join RiDiK and be part of an innovative, fast-growing team shaping the future of technology across industries.
We will review applications on a rolling basis until 27th Feb 2026 and early submissions are encouraged. Please note that only shortlisted candidates will be contacted. Thank you for your understanding.