Senior Agentic AI Engineer

Role overview

Responsibilities

• check_circle Design and build production-grade LLM and Agentic AI systems for planning, reasoning, tool use, long-horizon execution, and multimodal workflows.
• check_circle Develop coding agents, computer-use agents, multi-agent systems, and human-in-the-loop AI workflows.
• check_circle Build agent architectures involving planning, execution, state/memory management, structured tool calling, validation, retries, fallbacks, feedback loops, and deterministic tool integration.
• check_circle Integrate AI agents with APIs, databases, search systems, computational engines, enterprise applications, CAD/GIS platforms, and domain-specific software.
• check_circle Build production-grade agent harnesses and evaluation infrastructure for regression testing, trajectory evaluation, tool-call evaluation, model/prompt comparison, failure analysis, and release gating.
• check_circle Build RAG and knowledge systems using embeddings, vector/hybrid retrieval, reranking, metadata filtering, and grounded generation.
• check_circle Develop multimodal pipelines involving text, images, PDFs, scanned documents, technical drawings, maps, and structured data.
• check_circle Design and curate datasets for model training, fine-tuning, evaluation, retrieval, and synthetic-data generation.
• check_circle Fine-tune and adapt models using SFT, PEFT/LoRA/QLoRA, distillation, preference optimization, and other post-training methods where appropriate.
• check_circle Build AI observability and tracing for prompts, models, tool calls, agent trajectories, failures, retries, latency, token usage, and cost.
• check_circle Optimize AI systems for quality, reliability, latency, inference cost, and scalability.
• check_circle Translate cutting-edge AI research into reliable production systems.
• check_circle Lead technical design and architecture reviews and mentor junior and mid-level engineers.

Basic qualifications

• 5+ years of professional experience in AI/ML engineering, software engineering, or a closely related field.
• MS, PhD, or equivalent industry experience in Computer Science, Computer Engineering, Artificial Intelligence, Machine Learning, Applied Mathematics, or a related technical discipline; an interdisciplinary background combining Computer Science/Computer Engineering and Civil Engineering is highly preferred.
• Strong proficiency in Python, PyTorch, algorithms, data structures, debugging, profiling, testing, and systems design.
• Hands-on experience building and deploying production LLM and agentic AI systems.
• Strong understanding of LLMs, Transformers, prompt/context engineering, structured generation, tool/function calling, RAG, embeddings, and multimodal AI.
• Hands-on experience with coding agents, computer-use agents, multi-agent systems, and Model Context Protocol (MCP).
• Experience with Lang Graph, AutoGen/AG2, or equivalent custom agent orchestration infrastructure.
• Experience designing workflows involving planning, tool execution, state management, memory, retries, validation, feedback loops, and human-in-the-loop execution.
• Experience building LLM/agent evaluation harnesses, golden datasets, regression suites, trajectory/tool-call evaluations, tracing, observability, and production monitoring.
• Experience with model training/fine-tuning, dataset development, synthetic data, and model evaluation.
• Experience with RAG, vector databases, retrieval systems, and document-processing pipelines.
• Strong production engineering experience with APIs, databases, Docker, cloud infrastructure, CI/CD, and scalable systems.
• Ability to read, evaluate, reproduce, and adapt techniques from current AI research.
• Ability to independently take ambiguous AI problems from experimentation through reliable production deployment.
• Strong communication and technical leadership skills, including mentoring engineers and conducting architecture/code reviews.
• Experience or demonstrated understanding spanning Computer Science/Computer Engineering and Civil Engineering workflows is strongly preferred for this role, particularly where software or AI interacts with engineering calculations, spatial data, design systems, or engineering applications.

Preferred qualifications

• Publications at top-tier conferences such as NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, ACL, or EMNLP, or significant patents/open-source contributions.
• Experience with advanced post-training methods including SFT, LoRA/QLoRA, PEFT, DPO, RLHF/RLAIF concepts, continual pre-training, distillation, and synthetic-data training.
• Experience with Hugging Face Transformers, Datasets, Accelerate, PEFT, and TRL.
• Experience with inference optimization including KV-cache optimization, speculative decoding, quantization, batching, model routing, and latency/cost optimization.
• Strong LeetCode or competitive-programming background, particularly in graphs, geometry, optimization, search, dynamic programming, and data structures.
• Experience with cloud platforms, Kubernetes, distributed systems, FastAPI, PostgreSQL, PostGIS, and vector databases.
• Experience building AI systems for scientific, engineering, geospatial, or other technically demanding domains.
• CAD automation, engineering-software plugins, or AI agents capable of interacting with CAD/GIS/engineering applications.
• Experience translating civil-engineering workflows into deterministic algorithms, optimization systems, computational tools, or AI-agent capabilities