Hello, I'm
Aashish Gajadhane
Software Engineer
I build backend and AI systems that handle real production workloads.
3+ years building distributed backends, LLM/RAG systems, and AI pipelines. Recently scaled an AI media pipeline to 1,000 SKUs / 2,000+ images in ~30 minutes.
Focus Areas
Backend Systems
APIs, workers, databases, queues & performance.
AI / LLM Systems
RAG, vector search, AI pipelines & model integrations.
Distributed Processing
Async jobs, concurrency, retries & event-driven workflows.
I build backend systems and AI products.
I'm a Software Engineer working mostly with Node.js, Python, TypeScript, and AI/LLM systems.
Over the last few years, I've worked on everything from payment APIs and real-time backends to RAG systems and large-scale AI media processing.
Recently, I helped scale an AI media pipeline at CoreComAI from smaller catalog workloads to 1,000 SKUs and 2,000+ images in around 30 minutes.
At Troopr, I worked on LLM/RAG systems used by a platform serving 100K+ users, including vector search and Slack/Teams integrations.
I enjoy working on the backend-heavy parts of products where there's something interesting to figure out.
Beyond code
When I’m away from the keyboard, you’ll usually find me on a motorbike, riding through the ghats of Maharashtra or along the Konkan coast with my Wife🫰🏻 or the Boys.
Distributed Systems
Queues · Workers · Concurrency · Retries · Job orchestration
Production AI
LLMs · RAG · Vector Search · AI Pipelines · Evaluation
Backend Engineering
Node.js · Python · APIs · Databases · Event-driven systems
Engineering Experience
A few of the production systems I've worked on over the past few years.
Backend & AI Systems Consultant (Contract)
CurrentBuilt the backend architecture for an AI media generation platform that transforms product catalogs into marketing-ready images and videos. My work focused on distributed worker orchestration, GPU inference pipelines, asynchronous job execution, and designing a system that could scale reliably under long-running media workloads.
- Event-driven pipeline using FastAPI, ARQ, Redis, PostgreSQL, RunPod, and SSE.
- Dedicated FFmpeg worker pool separating CPU-bound rendering from GPU inference.
- Configurable concurrency, retries, and stage isolation for reliable execution.
- Structured logging and job lifecycle tracking across distributed workers.
Software Development Engineer
Worked on the backend systems powering an enterprise AI assistant platform used by 100k+ users. My work centered around knowledge retrieval, event-driven automation, and building internal tooling that enabled non-technical teams to manage enterprise documentation.
- Built Slack & Teams-triggered RAG workflows.
- Developed a multi-tenant Help Center platform with a real-time editor and live preview.
- Improved retrieval quality through prompt and search refinements.
- Integrated MongoDB Atlas Vector Search with LangChain and AWS Bedrock.
Full-Stack Engineer (Contract)
Built the media processing infrastructure behind a production social platform, focusing on adaptive video streaming, backend performance, and asynchronous processing.
- Event-driven HLS processing pipeline using S3 upload triggers and Lambda transcoders.
- Reduced API P95 latency by 30% via MongoDB aggregations, lean queries, and indexing.
- Increased API throughput by parallelizing independent queries with Promise.all().
- Delivered performance-critical backend features using Node.js, Docker, and AWS.
Backend Engineer
Designed and maintained core transactional backend APIs and payment infrastructure powering creator-business collaboration workflows.
- Integrated Stripe and PayPal payment gateways, handling complex webhook state transitions.
- Reduced Firebase read operations through batched retrieval strategies on high-traffic paths.
- Managed containerized backend deployments on AWS EC2 using Docker.
- Migrated backend codebase from JavaScript to TypeScript, enforcing strict contract typing.
Systems & Pipelines I've Shipped
Resume Coach
Distributed AI document processing pipeline with asynchronous workers, semantic vector matching, and real-time progress streaming.
Shramika NGO
Production-grade non-profit platform built with Next.js. Achieved 100/100 Lighthouse score, custom JSON-LD schemas for global SEO reach, and an integrated Razorpay donation flow that drove ₹15K+ in organic international contributions within 7 days of launch.
Event-Driven Media Transcoding Architecture
Event-driven, serverless media pipeline for adaptive bitrate video streaming. Built based on the production system deployed at SEEme Inc to automate high-throughput transcoding.
Technical Deep Dives
Architecture breakdowns, engineering decisions, and lessons learned while building production systems.
Designing a Distributed AI Media Processing Pipeline
An architectural deep dive into building CoreComAI's production media pipeline: why ARQ was chosen over BullMQ for Python workers, how decoupling CPU-bound video stitching from GPU inference solved starvation bottlenecks, 15-item batching strategies, and real-time SSE progress streaming.
How I Found a Fake Job Assessment Repo Hiding Malware Inside SVG Files
While auditing a take-home interview assignment, I uncovered malware hidden inside static SVG files. This write-up walks through how the exploit was structured, what it targets, and how to safely audit third-party code before running it locally.
Building Telegram Bots with Telegraf.js
A practical guide to building Telegram bots with Telegraf.js, covering architecture, environment configuration in Node.js.
The tools I build with
Technologies and systems I use to build scalable, production-ready applications.
Distributed Systems & Backend
Architecting reliable systems for scale, performance, and real-time workloads.
AI & Machine Learning Infrastructure
Building production-ready AI workflows and LLM-powered applications.
Data & Caching
Designing data layers optimized for speed and reliability.
Infrastructure & DevOps
Deployments, automation, and cloud infrastructure at scale.
Let's build something together
Interested in backend engineering, AI infrastructure, distributed systems, or event-driven architectures? I'd love to chat.
Open to full-time roles and high-impact projects
aashishdg.work@gmail.com
GitHub
github.com/arsen1c
aarsen1c
aashish-gajadhane