AI Neuroscience Intern Obten
· Remote (India preferred, English preferred, geo agnostic)
· Internship What Obten is An AI-native B2B export marketplace connecting Indian manufacturers with global buyers. Small team, moves fast, AI-first from day one. The role We run a proprietary AI system at the core of our operations. Your job: monitor its health, spot where it drifts or miscalibrates, and help us keep it sharp. Think site reliability for a reasoning system. This is a research and judgment role. You would be working at the leading edge of neuroscience x AI — not on toy problems or benchmarks, but on a live system making real business decisions every day.
What you'll do
• Analyse AI decision logs and find patterns — recurring errors, confident-but-wrong outputs, wasted cycles
• Track which instructions are followed vs. ignored
• Build a weekly health report and propose fixes What you should know
• How LLMs process and lose context across a session — working memory limits and attention degradation
• How instruction-following breaks down: compliance vs. genuine reasoning
• Confidence miscalibration — fluent, plausible outputs that are wrong,
and how to detect them
• Cognitive load analogues: how prompt length, task switching, and context accumulation affect output quality Who we're looking for Creative polymaths. People who move between cognitive science, philosophy of mind, linguistics, systems thinking, and AI — and find the connections more interesting than any single field. Biology students with a strong interest in AI and the drive to learn fast are very welcome. What matters more than your background: you read quickly, think on your feet, pick up new frameworks without hand-holding, and move quick. Please share your academic scores, any test results (GRE, GATE, competitive exams, whatever you have), and something you've read recently that changed how you think. Why it's unusual You'd be studying an AI system running a real business, in production, with real stakes. The questions you'd be working on — how reasoning degrades, how confidence miscalibrates, how instruction-following breaks down — don't have settled answers yet. You'd be helping write them.
📌 AI Neuroscience Intern (Manesar)
🏢 Obten
📍 Manesar