iMerit Technology

An AI agent’s exploit score only means something if the sandbox holds. In ExploitGym evaluations at two frontier labs, agents reached real infrastructure, and some kept going after noticing the target looked real. imerit.ai/resources/blog/wha… #AgentEvaluation #AISafety

A busy week for the iMerit team at @_odsc West 2026! From conversations with the AI community to Brian Smith’s session on building hard benchmarks with academic experts, it was a week of sharing ideas on how to make AI evaluation more rigorous & reliable. #ODSCWest #AIEvaluation

Radha Basu is featured in @IndiasporaForum's 250 at 250, honoring Indian American leaders shaping business and society. Her journey spans entrepreneurship, AI, and expanding tech careers through Anudip Foundation. List: 250moments.indiaspora.org/ #Indiaspora250 #Leadership #iMerit

iMerit, an EXL company, launches iMerit Scholars, powered by Ango Hub, to manage 30,000+ experts across 118 countries and 106 languages for AI model training & evaluation. Its first major launch since @exl_service acquired iMerit in July. globenewswire.com/news-relea… #iMeritScholars

Join Tiffany Thompson at Vehicle Tech Week North America in Novi, Michigan. Her session covers how teams can manage human decisions as AI programs grow. Wednesday, October 28 | 11:50am ET Register here: register.visitcloud.com/surv… #VehicleTechWeek #AutonomousVehicles #HiTL

What does it take to prepare data for frontier model training? From off-the-shelf data to production-ready training inputs, explore the pipeline that helps improve model performance. Read the blog: imerit.ai/resources/blog/fro… #AI #DataQuality #OTS #AITraining

How can academic experts help build AI benchmarks that challenge even top models? Brian Smith of iMerit, an EXL company, explores this at @_odsc WEST 2026, sharing lessons from a phonology evaluation. Oct 27 | 11:05–11:35 AM #AI #AIEvaluation #ODSC

An autonomous tractor that sees perfectly at noon can go blind at dusk or in heavy rain. Low light, dust, glare, and bad weather are where perception fails. The fix is not more images. It is training data that covers these hard conditions. imerit.ai/resources/blog/aut…

We are at ODSC AI West 2026. Come meet the iMerit team at Booth 6 to talk training data and evaluation pipelines for frontier AI, from computer vision to model evaluation. October 27 to 29, Burlingame, CA. Stop by and say hello. #ODSCWest #GenerativeAI #iMerit

AI video looks convincing for a few seconds, then a face shifts or an object jumps between frames. That is temporal drift, and it becomes a real problem in production. Fixing it takes sequence level evaluation, better training data, and expert review. imerit.ai/resources/blog/sol…

Siddhartha Bal, Director of Autonomous Mobility at iMerit, joins the panel “Redefining Value Through Intelligence, Innovation and Outcomes” at the Nasscom Technology Confluence 2026 in Coimbatore, with leaders from Bosch, TCS, and DesiCrew. #nasscom #iMerit

AMR autonomy gets hard with sensor noise, localization drift, mixed terrain, and rare obstacles datasets miss. iMerit structures camera, LiDAR, and robot logs around the classes and safety zones AMR models plan with. imerit.ai/domains/robotics/a… #AutonomousMobileRobots #Robotics

At @TechCrunch Disrupt 2026, Radha Basu, EVP & Head of iMerit at EXL, will lead the roundtable "Arm Wrestling Your Data: The Physical AI Challenge" on data quality for physical AI. Roundtable 1 | Oct 13 | 10:00–10:30 AM 🔗: techcrunch.com/events/techcr… #PhysicalAI #TechCrunchDisrupt

AV data pipelines are shifting to directed triage, where AI flags model weaknesses and requests targeted scenarios. A model can't verify the root cause or confirm new data fits the test. Learn how human experts validate it: imerit.ai/resources/blog/whe… #AutonomousVehicles #AVData

Erikk Cass, AVP, Mobility, iMerit, an EXL company, told @Forbes why humanoid robot safety has less to do with smarter models and more to do with how quickly a robot learns from what it gets wrong. Read here: forbes.com/sites/johnkoetsie… #HumanoidRobots #PhysicalAI #Forbes

Radha Basu, EVP & Head of iMerit at EXL, joins @TechCrunch Disrupt 2026 in San Francisco, October 13 | 10:00–10:30 AM She’ll share how physical AI can move from promising pilots to reliable systems in the real world. Learn more: techcrunch.com/events/techcr… #PhysicalAI #Robotics

AV expansion creates coverage gaps. Map territory differences, rank scenarios by risk & route gaps to annotation, simulation, or validation. See how AV teams build smarter triage programs for new territories: imerit.ai/resources/blog/how… #AI #DataTriage #SelfDrivingCars #AV

Ask five frontier labs for their HumanEval scores and you’ll get nearly identical numbers. That’s not differentiation. It’s benchmark saturation. See how labs are building fresh, expert-reviewed evaluations from production data: imerit.ai/resources/blog/wha… #LLMEvaluation #GenAI

Open-field navigation models don't survive orchard headlands. Tight turns, tree obstacles, and weak GPS demand orchard-specific trajectory labeling across crop types and conditions. Read the blog to learn more: imerit.ai/resources/blog/why… #AgriculturalAI #AutonomousMobility

iMerit built a platform to track clinical trial data at scale: versioning changes, resolving trials across sources & flagging uncertain data. ◼️ 1,424 linkages ◼️ 727 entities resolved ◼️ 357 revisions tracked. Case Study: imerit.ai/resources/case-stu… #ClinicalTrials #LifeSciences