Aman Vijay | Full Stack Software Engineer

Aman Vijay · Full Stack Software Engineer

I build full-stack products that

don't fall over.

Full Stack Software Engineer — Next.js and React on the front, FastAPI and Node on the back, RAG pipelines and AWS/GCP in production. 1+ year shipping web apps, AI features, and event-driven backends. Currently Junior Software Developer at Psypher AI (Outyn).

view my work →

amanvijaywork@gmail.com

live signals

stack
Next · FastAPI · Redis · AWS
Full stack
role
Junior Software Developer · Outyn
Psypher AI
shipping
events · semantic search · itinerary
Outyn

Correctness first. Then performance. Elegance is a distant third — but it matters.

scroll

Background

I'm a Full Stack Software Engineer — currently a Junior Software Developer at Psypher AI (Aug 2025–Present), building Outyn: event-driven notifications on Redis Streams, async member invites, hybrid semantic search, and AI itinerary planning with Redis caching.

Before that I was a Full Stack Developer Intern at InnovateWise Solutions (Jan–Dec 2024): led interns on AdviceBytes, shipped a MERN cybersecurity platform, JWT + Google OAuth, a JSON policy form builder, and CI/CD to AWS EC2 via GitHub Actions.

B.Tech CSE, JK Lakshmipat University (July 2024). Side projects: RepoLens and AskYourDocs — retrieval pipelines, Postgres/pgvector, and FastAPI backends.

TypeScript
React
Next.js
FastAPI
Node.js
PostgreSQL
Redis
MongoDB
OpenAI
LangChain
Docker
AWS
GCP

GitHub ↗LinkedIn ↗X ↗Also: films, anime, lifts →

1+ yr

shipping production web · AI · cloud

Full stack

Next · FastAPI · Postgres · Redis

GCP

Associate Cloud Engineer · 2023

B.Tech CSE

JK Lakshmipat · July 2024

From the notification worker at Psypher AI — a real request-latency fix. Sequential awaits were stacking cost on every search; the bottleneck was the wait, not the compute.

# run concurrently — saves ~600ms per request
result, parsed = await asyncio.gather(
    moderate(query),
    llm_parse(query)
)

Skills

Click a skill to filter projects. Click again to clear.

Frontend

Backend

AI / Data

Cloud

Experience

aman@portfolio ~/experience

$ git log --author="aman@vijay" --format="%h %s" --stat

a3f91c2

[psypher-ai]shipped Outyn event platform + Redis Streams notifications

Outyn · Redis Streams · async invites | event-driven services

Aug 2025 – Present

d12c3e1

[psypher-ai]hybrid semantic search + AI itinerary with Redis cache

vector search · LLM query understanding · Redis caching

Aug 2025 – Present

b8f21a0

[innovatewise]led AdviceBytes + MERN cybersecurity / AI learning stack

React · Tailwind · MERN · OpenAI GPT | Jan–Dec 2024

Jan 2024 – Dec 2024

3c910f4

[innovatewise]JWT, OAuth, policy builder, and AWS CI/CD

JWT · Passport · Google OAuth · GitHub Actions · AWS EC2 · Docker · NGINX

Jan 2024 – Dec 2024

Projects

Featured

RepoLens

Helps engineers understand unfamiliar codebases: import a repo, get a structured map, chat over it, and plan implementations — never generate code.

Problem — Dropping into a strange monorepo means hours of grep before you can propose a safe change.

Why shared packages/db, not api→worker imports — API and workers stay peers. Both depend on packages/db so the worker never becomes a sidecar of the HTTP layer.
Why ARQ, not Celery — Backend is async-first. ARQ is asyncio-native; Celery’s async story is bolted on and heavier than this queue needs.
Why Neon + Upstash from day one — Managed Postgres/pgvector and Redis kill local dual-config drift. Dev and prod speak the same wire protocol.

Case study GitHub Live