AfterQuery

The best model for finance isn't the one with the highest benchmark score. It's the one you can afford to run across the whole portfolio. A single loan can be governed by a base agreement and a series of amendments. A portfolio can mean thousands of documents. At that volume, cost per document matters as much as accuracy. With Hebbia, we tested whether expert data could close the gap. We post-trained NVIDIA’s Nemotron 3.5 Super VL on finance tasks written by former credit investors and our in-house experts. The result: - Perfect solves rose from 47% to 74% - Quotes instead of paraphrasing - Within 3 points of Astra - About 1/7 the cost Teams can run post-trained open model on the high-volume work and route edge cases to a frontier model. Thanks to Hebbia and NVIDIA. Full case study in the comments.

The AfterQuery and NVIDIA teams are hosting a happy hour tonight at 6:30pm for SF Tech Week. If you’re working on AI research, post-training, or enterprise AI deployment, join us for a drink and to meet the people pushing the frontier. Drop your email below and we’ll send you the invite!

AfterQuery is teaming up with Goldman Sachs to host The Frontier of Finance at SF Tech Week. Every knowledge work industry will be reshaped by AI. Finance will be one of the first. On October 7, from 6-8:30 PM, we'll dig into AI implications in finance and how the industry is adapting. If you work in finance, research, or AI, join us. Stay after the panel for food, drinks, and time with the speakers. Drop your email below and we’ll send you an invite!

Excited to welcome Charles Hoff as an Engagement Manager at AfterQuery! Charles joins us from Surge AI, where he led customer engagements end-to-end. Before moving into tech, Charles practiced corporate law at a major firm and clerked for a federal judge. Now Charles is joining AfterQuery to own frontier lab engagements, understanding what each lab actually needs and making sure AfterQuery’s work answers it. If you want to join a world-class team, check out our open roles in the comments.

Excited to welcome Kayla (DeFrank) McElwee as AfterQuery’s AI Safety & Security Lead! Kayla most recently founded and led the AI safety and red teaming practice at a human data startup, after standing up the same capability at another. Before moving into tech, Kayla spent five years as an intelligence officer, supporting defense planning, operations, and crisis response. Now Kayla's joining AfterQuery to lead our Safety & Security business and set the bar for the evaluations our customers depend on. If you want to work with Kayla and people like her, check out our open roles in the comments.

Excited to welcome Kayla (DeFrank) McElwee as AfterQuery’s AI Safety & Security Lead! Kayla most recently founded and led the AI safety and red teaming practice at a human data startup, after standing up the same capability at another. Before moving into tech, Kayla spent five years as an intelligence officer, supporting defense planning, operations, and crisis response. Now Kayla's joining AfterQuery to lead our Safety & Security business and set the bar for the evaluations our customers depend on. If you want to work with Kayla and people like her, check out our open roles in the comments.

Congratulations to the Motif Technologies team on the release of Motif 3! AfterQuery is proud to have served as the sole data partner on this model. This is a historic moment for Korean sovereign AI and the open weights ecosystem writ large. We look forward to supporting Motif and the Korean AI ecosystem as they push the frontier of open intelligence. Motif 3 leads its comparison set on τ³-Banking, and posts 94.7 on τ²-Bench Telecom, 74.9 on Terminal-Bench 2.1, and 76.2 on SWE-bench Verified. It’s also great to see NVIDIA’s open foundation put to use here - Motif 3's post-training was done via NeMo-RL. In a world where it’s easy to assume that the AI race is already won, the team at Motif is proof that the field is open, and we’re still at the starting line.

Congrats to the Legora team on the release of the Legora BAR, a benchmark for agentic legal work! BAR runs on 5,161 real law firm cases across 28 practice areas, graded against expert-verified rubrics. AfterQuery helped QA the benchmark with Legora. Proud to work with their team on measuring the frontier of agentic legal work. Link to the full blog post in the comments. 

To the researchers impacted by the recent Amazon AGI layoffs: AfterQuery is hiring. We're looking for post-training researchers who care deeply about improving frontier model capabilities. At AfterQuery, you'll help tackle some of the hardest problems in the space and run experiments that support teams across all frontier AI labs. DMs open, or apply at the link in the comments.

Last Thursday, we hosted Poker Night at our SF office. Researchers, founders, and builders joined the AfterQuery team for 3 things: 1. Poker. 2. Drinks. 3. Lively conversation. We're hosting more events soon. Comment if you want to join the next one ↓