Advay Sinha
backend developer · ai/ml engineer
I’m a Computer Science student and backend-focused developer passionate about building scalable systems, AI-driven applications, and high-performance web platforms. My work spans backend engineering, network monitoring platforms, speech recognition integrations, algorithmic trading simulations, and full-stack development using TypeScript, Node.js, React, MongoDB, and cloud technologies.
Project Vault
Network Intrusion Monitoring System
OPERATIONALopen dossier →Watches a critical production network it is structurally incapable of modifying — no write path to any device exists in the codebase.
Multi-Agentic Research Platform
OPERATIONALopen dossier →Answers research questions only with claims it can verify — every stage traced, every citation grounded.
Algo Trade Simulator
OPERATIONALopen dossier →Lets strategies face real market data without letting simulated capital pretend to be real.
FloatChat
OPERATIONALopen dossier →Makes a heavyweight scientific archive answerable in plain language, from a cached store that fits on one machine.
Mock AI
OPERATIONALopen dossier →Generates and grades interviews with an LLM while keeping auth, sessions, and scoring accountable to a real backend.
Journey
Schooling upto X
Kairali School, Ranchi
▸ CBSE board score: 94%.
Senior secondary(XII) — Science (maths)
Delhi Public School, Ranchi
▸ CBSE board score: 93%.
B.Tech, Computer Science and Engineering — AI specialization
Bennett University, Greater Noida
▸ in progress · cgpa 8.7 / 10.
Project management, junior core
Enactus, Bennett University
One tenure of project coordination across social-enterprise builds — scope, schedules, and people instead of endpoints.
Backend software developer — speech recognition & safety infrastructure
Zarva, Greater Noida
Engineered a real-time speech recognition pipeline and 5+ REST APIs over WebSockets for live call connectivity — instant distress detection across 100+ concurrent cab rides. Built a graph-theoretic shortest-path routing algorithm over integrated crime datasets, cutting risk-exposure scores 40% in high-incident zones.
▸ shipped to production; first engineering role.
1st place among 40+ teams
Mobile Next — MOBILON App Showcase
Won on Zarva's real-time cab-safety system — WebSocket call streaming, live distress detection, risk-aware routing — demonstrated to a panel of industry judges.
SIH 2025 — pre-national selection stage
Smart India Hackathon — Bennett University
Advanced to the pre-national selection stage among 450+ competing teams; certificate of achievement on record.
AI data & evaluation contractor — frontier model alignment
Alignerr (freelance, remote)
Evaluated and labelled large-scale model-output datasets for frontier AI systems (clients including Anthropic) against structured rubrics — 10+ evaluation cycles, plus adversarial test cases probing logical and algorithmic edge cases across code generation, reasoning, and multi-step instruction-following.
▸ contract closed.
Software engineering intern — networks & cybersecurity
Airports Authority of India — Jay Prakash Narayan International Airport, Patna
Architected and deployed NIMS: ML intrusion detection, telemetry health prediction, and configuration auditing unified across the airport’s multi-vendor campus network (Huawei, Cisco, Hirschmann, Sophos). Field-validated a read-only ingestion layer — SNMPv3 polling, traps, syslog, SSH, SIEM API — against a production Huawei S5720, surfacing 87 configuration findings with rollback-verified dry-run remediation.
▸ read-only by construction: no code path can modify a device.