Aditya Sinha

Backend engineer with a track record of building reliable data and media pipelines, reducing operational costs through infrastructure and architecture optimization, and shipping backend systems with production-grade reliability.

Currently at Skidos, I own an end-to-end video ingestion pipeline transcoding IP content into HLS adaptive bitrate streams and a multi-step Voice-over pipeline deployed across 14 languages. Previously at Samespace, I built high-throughput notification delivery systems and AI-driven semantic search infrastructure on QdrantDB.

Outside of work, I enjoy gaming and I love badminton — I used to play at a national level.

Experience
SkidosSoftware Engineer● Current
March 2025 – Present
  • ○HLS Pipeline: Owned end-to-end video ingestion transcoding IP content into chunked HLS adaptive bitrate streams, enabling SDK playback and reducing delivery bandwidth by serving segments on demand.
  • ○Voice-over Pipeline: Used GPT-4 to localize product content across 14 languages via a kid-centric rule engine; synthesized audio with ElevenLabs SDK, cutting turnaround from days to minutes.
  • ○DB Cost Reduction: Migrated stale data to cold storage and overhauled indexing and retention strategy; improved query performance across hot-path APIs.
  • ○API Refactor: Led phased refactor of 150+ APIs into Domain-Driven Design service layers, coordinating with QA and platform teams to reduce tech debt without impacting production latency.
SamespaceSDE-1 Backend
March 2024 – March 2025
  • ○Notification System: Built a high-throughput notification delivery system; reduced P99 latency by 80% through Edge Caching, improving reliability at scale.
  • ○AI Search: Developed AI-driven contact and conversation search on QdrantDB, enabling both semantic and full-text retrieval via document captioning and chunking.
  • ○LLM Infrastructure: Built backend for document-grounded LLM conversations with gRPC inter-service communication, QdrantDB vector retrieval with cosine similarity and confidence thresholding.
  • ○Observability: Implemented production observability stack with SigNoz and OpenTelemetry for real-time performance insights, alerting, and pipeline health monitoring.
Skills
Go
Docker
AWS
Kafka
RabbitMQ
MongoDB
Redis
MySQL
JSJavaScript
Node.js
Google ADK
Claude
Projects
HLS PipelineTrading Agent
Links
Last Played
—
Not available
live
😊
Claude Usage
all good
resets in
1h 30m
2hr session25% used
0m30m1h1.5h2h
this week
M
T
W
T
F
S
S