ML/AI Engineer | Tenstorrent Global Negotiation Guide

Negotiation DNA: [Pre-IPO Equity + Base + Sovereign AI Bonus] | RISC-V AI Silicon | Blackhole/Wormhole Architecture | Jim Keller Leadership | +15-25% AI Premium

Region	Base Salary	Equity (Options/RSU est.)	Bonus	Total Comp
Toronto	C$165K–C$218K	C$62K–C$105K	12–18%	C$255K–C$358K
Austin	$155K–$208K	$50K–$82K	12–18%	$235K–$325K
Remote US	$148K–$202K	$48K–$78K	12–18%	$222K–$312K

Negotiation DNA

ML/AI Engineers at Tenstorrent optimize the intersection of machine learning workloads and RISC-V AI silicon — developing model compilation pipelines, kernel implementations, and AI framework integrations that demonstrate Blackhole and Wormhole chips can match or exceed NVIDIA's performance on transformer, diffusion, and LLM workloads. Under Jim Keller's leadership, ML/AI Engineers are critical to proving that open-architecture RISC-V silicon can compete with NVIDIA's proprietary CUDA ecosystem on the workloads that matter most to Sovereign AI customers. The +15-25% AI Premium reflects the explosive demand for engineers who bridge ML algorithms and custom silicon. At a late-stage startup valued at ~$3.5B, ML/AI Engineers shape the product-market narrative. (Sources: Tenstorrent ML engineering postings, 2025; AI engineer compensation surveys, 2025–2026.)

Level Mapping: Tenstorrent ML/AI Engineer ~ Google L4–L5 ML | Meta ML E4–E5 | Apple ML IC3–IC4 | AMD AI Software Engineer | NVIDIA CUDA Engineer | DeepMind Research Engineer

RISC-V AI Silicon — The Sovereign AI Premium

ML/AI Engineers at Tenstorrent are at the absolute center of the Sovereign AI value proposition. When a nation evaluates whether to build domestic AI capability on Tenstorrent's RISC-V chips versus NVIDIA, the deciding factor is AI workload performance — can Blackhole and Wormhole silicon run LLMs, diffusion models, and transformer inference at competitive performance-per-watt and performance-per-dollar? ML/AI Engineers directly determine this outcome by optimizing model compilation, writing high-performance kernels, and building the AI framework integration layer (PyTorch, JAX, TensorFlow) that makes Tenstorrent silicon accessible to Sovereign AI developers. Dozens of nations — the EU, Japan, India, Canada, Saudi Arabia, UAE, South Korea, and more — are investing billions in AI independence, and every one of them needs proof that open-architecture silicon can run their workloads. Jim Keller's track record (AMD Zen, Apple A-series, Tesla FSD) ensures the silicon is world-class; ML/AI Engineers make it sing for AI workloads. The +15-25% AI Premium is justified by the extreme scarcity of engineers who combine deep ML expertise with silicon-level optimization. Candidates should negotiate for Sovereign AI performance bonuses of $18K–$28K annually, tied to benchmark results and Sovereign AI customer adoption milestones.

Global Levers

1. AI Premium Equity Acceleration: "The +15-25% AI Premium applies to equity as well as base. Given the ~$3.5B valuation and the fact that my ML optimization work directly determines whether Sovereign AI customers adopt Tenstorrent, I'm requesting a 30% equity uplift with performance-accelerated vesting tied to benchmark milestones."
2. Sovereign AI Performance Bonus: "My kernel and compiler optimization work directly determines whether Sovereign AI nations choose Tenstorrent over NVIDIA. I'm requesting a $20K–$28K annual Sovereign AI performance bonus tied to MLPerf results, customer benchmark wins, and Sovereign AI adoption metrics."
3. CUDA-Competitive Talent Scarcity: "Engineers who can write high-performance AI kernels for custom silicon — competing directly against NVIDIA's decade-long CUDA investment — are among the rarest talent in the industry. My experience with [custom AI kernel development/MLIR compiler optimization/hardware-aware ML] justifies a base salary of $202K+ and a $25K–$35K signing bonus."
4. Research Publication and Conference Budget: "Demonstrating Tenstorrent's AI performance at top venues (MLSys, ISCA, NeurIPS) is critical for Sovereign AI customer credibility. I'm requesting a $10K annual research budget and conference travel allowance."

Negotiate Up Strategy: "I'm deeply excited about optimizing AI workloads for Tenstorrent's RISC-V silicon under Jim Keller's leadership — this is the frontier of AI hardware-software co-design. I have competing offers from [NVIDIA at $305K / Google at $298K / Meta at $295K] total comp. Given Tenstorrent's pre-IPO equity at ~$3.5B, the +15-25% AI Premium, and the Sovereign AI tailwind, I'm targeting $305K–$312K total comp: $202K base, $78K equity, and a $25K Sovereign AI performance bonus. My accept-at floor is $282K total comp — below that, the risk-adjusted value against NVIDIA's liquid comp doesn't clear. I can sign within 72 hours at $295K+ and immediately begin optimizing LLM inference on Wormhole silicon. The pre-IPO equity upside and Jim Keller's track record make Tenstorrent the most compelling opportunity in AI hardware."

Evidence & Sources

• Levels.fyi ML/AI Engineer compensation at NVIDIA, Google, Meta, and semiconductor companies, 2025–2026
• AI engineer salary surveys (ai-jobs.net, Rora negotiation data), 2025–2026
• Tenstorrent ML engineering and AI performance job postings, 2025
• Sovereign AI workload requirements and MLPerf benchmark analysis, 2025–2026