Senior ASIC Design Engineer — AI Evaluation & Benchmarks

We’re looking for an experienced chip design engineer to evaluate Mistral models on real-world semiconductor design workflows. You’ll combine hands-on engineering expertise with benchmark development and technical review to improve how AI models perform tasks across RTL development, synthesis, physical design, and tapeout.

What You’ll Do

• Evaluate AI model performance: Assess recent Mistral models on chip design tasks, identifying errors, limitations, and opportunities to improve technical accuracy and practical usefulness.
• Develop chip design benchmarks: Scope realistic tasks with clear requirements, appropriate difficulty levels, reference solutions, and measurable evaluation criteria.
• Review vendor-delivered tasks: Validate tasks involving RTL coding, logic synthesis, and synthesized netlists for correctness, completeness, reproducibility, and relevance to engineering workflows.
• Guide synthetic task generation: Provide domain expertise across the design flow, from RTL coding and verification through physical design, timing closure, and tapeout.
• Create high-quality seed tasks: Develop original chip design problems, supporting artifacts, and reference solutions to seed synthetic task generation.
• Define quality standards: Establish review criteria that account for functional correctness, synthesizability, timing constraints, and power, performance, and area.
• Collaborate with AI and data teams: Translate engineering findings into actionable feedback that improves model evaluations and generated training data.

What You Bring

• Significant hands-on experience in ASIC or SoC design, with deep expertise in RTL development and synthesis.
• Strong proficiency in Verilog or SystemVerilog, including writing, debugging, and reviewing synthesizable RTL.
• Experience with EDA tools for simulation, synthesis, and implementation, including interpreting reports and diagnosing design issues.
• Working knowledge of the RTL-to-GDSII flow, including static timing analysis, physical design, timing closure, and signoff.
• Ability to create self-contained engineering tasks with clear specifications, testbenches or validation methods, and technically sound reference solutions.
• Strong technical judgment and attention to detail when reviewing code, netlists, constraints, and design artifacts.
• Clear written communication and the ability to explain complex design concepts to AI researchers, engineers, and data teams.

Preferred Qualifications

• Experience contributing to a chip design through tapeout.
• Experience evaluating AI-generated code or developing technical benchmarks, datasets, or automated validation workflows.
• Proficiency in Python, Tcl, or shell scripting for EDA automation and results analysis.
• Familiarity with formal verification, equivalence checking, or open-source chip design tools.