{"$schema":"https://raw.githubusercontent.com/jsonresume/resume-schema/v1.0.0/schema.json","basics":{"name":"Anurag Akkiraju","label":"Software Development Engineer II","summary":"AI systems engineer building long-horizon agent runtimes and LLM platforms at Amazon, with expertise in prompt caching, multi-agent pipelines, and evaluation harnesses.","location":{"city":"Seattle, WA"},"profiles":[{"network":"LinkedIn","username":"anurag-akkiraju","url":"https://linkedin.com/in/anurag-akkiraju"},{"network":"GitHub","username":"maskedband1t","url":"https://github.com/maskedband1t"}]},"meta":{"canonical":"https://saywise.com/member6132","version":"v1.0.0","lastModified":"2026-10-05T19:47:07.521Z"},"x_saywise":{"handle":"member6132","pronouns":null,"availability":null,"oneLiner":"AI systems engineer building long-horizon agent runtimes and LLM platforms at Amazon, with expertise in prompt caching, multi-agent pipelines, and evaluation harnesses.","profileUrl":"https://saywise.com/member6132","markdownUrl":"https://saywise.com/member6132/profile.md","pdfUrl":"https://saywise.com/member6132/resume.pdf","archetype":{"code":"RS-MX","name":"Frontier Artisan","url":"https://saywise.com/sca/694794fa0070","rarity":"One of the first Frontier Artisans on Saywise"}},"work":[{"name":"Amazon","position":"Software Development Engineer II","location":"Seattle, WA","url":"https://amazon.com","startDate":"2022-08-01","endDate":"2024-07-01","summary":"Building next-generation personal AI agents, owning the runtime for long-horizon durable work and the proactive layer that listens for real-world changes. Led the agent's prompt-caching program, improving cache hit rate from 45.6% to 83.0% while reducing latency by 57%. Led the Alexa+ notification domain rebuild, designing the model interface and evaluation harness for ~90M customers. Designed a platform primitive for deterministic service actions and contributed to the real-time proactive content system, which achieved 26x engagement improvement."},{"name":"Lawrence Livermore National Laboratory","position":"Machine Learning Research Intern","location":"Livermore, CA","startDate":"2021-06-01","endDate":"2021-08-01","summary":"Designed and trained an LSTM in PyTorch for sequence labeling over raw binary data, classifying non-code byte regions in PE and ELF binaries at 95%+ accuracy. Built an approximate nearest-neighbor retrieval system for software origin tracing using NMSLIB, returning k-nearest neighbors for newly ingested binaries against an enterprise corpus. Extended the work through spring 2022."}],"education":[{"institution":"University of Florida, Herbert Wertheim College of Engineering","studyType":"B.S. Computer Science, Minor in Electrical Engineering"}],"projects":[{"name":"Calibrated Decisions at the Human–Robot Boundary","description":"Investigated the decision layer between a robot's planner and its policy (act, ask, or hand off to a person). Tested calibrated models against frozen rules and an oracle on four simulated bodies across 160+ experiments. Distilled a 421M-parameter on-device model from a cloud judge, maintaining 88.3% accuracy versus 87.9% for the cloud version at 90ms latency with calibration intact. Identified a failure mode in fleet learning where retraining on operator takeovers can make models confidently wrong outside corrected regions."}],"skills":[{"name":"long-horizon agent runtimes"},{"name":"agent memory"},{"name":"multi-agent LLM pipelines"},{"name":"prompt caching"},{"name":"PyTorch"},{"name":"reinforcement learning"},{"name":"calibration"},{"name":"hybrid FTS5/dense retrieval"},{"name":"evaluation-harness design"},{"name":"instruction specs"},{"name":"frozen baselines"},{"name":"regression testing"},{"name":"non-inferiority testing"},{"name":"failure-mode analysis"},{"name":"Python"},{"name":"Java"},{"name":"C++"},{"name":"TypeScript"},{"name":"SQL"},{"name":"Bash"},{"name":"AWS"},{"name":"Docker"},{"name":"Git"},{"name":"CI/CD"},{"name":"distributed systems"},{"name":"event-driven pipelines"}]}