Backend engineer at Flashback Labs. Karimnagar, India.
I work on backend and infrastructure — Node and Go services, Postgres, and SQS-driven pipelines on AWS. Most of what I build at work is in private repos, so briefly, what it is:
ImageShield — likeness-monitoring platform. Node/TypeScript/Fastify and Python/FastAPI over Postgres 16, across 10 containerised services on AWS ECS. Eliminated IDOR across 103 HTTP routes with a build-failing test driven off the live route table. 74 reversible migrations gated in CI by an apply, rollback and diff pipeline.
Legacy Mode — the Node/Express backend behind a production AI product, built across 23 phases. Cut voice first-audio latency about 10x after benchmarking four TTS vendors on latency and cost. Cross-process realtime delivery over Postgres LISTEN/NOTIFY rather than adding a pub/sub dependency.
Media pipeline — Go and TypeScript ingestion on SQS, DynamoDB and S3. Bounded-concurrency Go workers running ONNX face detection, and a three-tier dedup stage that removed 34.5% of jobs as URL duplicates and another 4.6% as exact content duplicates.
AutoDev — browser-based AI code editor. Runs a full Node runtime in the browser through WebContainers, with agents across Gemini, OpenAI and Groq. Built Jan–Mar 2026; mine apart from six commits from a collaborator.
Portfolio — source of mokshith.vercel.app.
CardioGuard — cardiac arrest detection from real-time ECG signals. Built with a team; won at a hackathon.
Qconn — Q&A platform for campus knowledge sharing.
TypeScript, Go, Python, SQL · Node.js, Express, Fastify, FastAPI · PostgreSQL, DynamoDB · AWS (ECS, SQS, S3, IAM, Secrets Manager), Terraform, Docker · React, Next.js