Founding engineer.
Product builder.
AI systems thinker.
Co-founded a startup. Shipped systems under real load. Now going deeper on AI — at Northeastern and in production.
The Story
The Story So Far
From a C compiler to a co-founded startup to grad school. The path was never straight — but it was always moving forward.
First Line of Code
A C compiler, a blank terminal, and a loop that printed my name. Small moment. Total redirection. The language forced me to think about memory — about what the machine is actually doing. That foundation never left.
gcc hello.cBuilding Worlds with Java
Moved into Java and built games — including Hunt the Wumpus. For the first time, software felt like a world I'd designed: rules, state, consequences. That shift in mental model — from code as instructions to code as systems — is still how I think.
Hunt the WumpusFirst Invoice
Started freelancing. Real clients, real deadlines, real consequences for bad code. Working on actual production problems — not toy exercises — taught me more in twelve months than the previous two years combined. Also learned what 'done' actually means.
First paid projectCodrel — Building a Community
Started Codrel, a community for people learning to code. First time I built something for a community rather than a client. Learned that product decisions when users are real — not hypothetical — hit completely differently.
Codrel platformCo-founded Dcrypt
Launched an AI-powered coding education startup. Built a microservices architecture, a containerized remote code execution engine running Python, JavaScript, Java, C, and C++ in Docker sandboxes, and async job pipelines on RabbitMQ. Ran it in production for two years — 99.9% uptime. The 3am deployment fires taught me things no course ever could.
Production system · 99.9% uptimeResearch @ IIIT-H · Smart Cities
Software Research Intern at IIIT-H under the Government of India's Smart City Lab. Built backend APIs and computer vision pipelines — YOLOv5 inference, 200+ concurrent camera feeds, 10K+ daily commuters monitored in real time. Infrastructure for real operational load.
Smart City Lab · Govt of IndiaBachelor's — Chapter Closed
Graduated with my bachelor's degree. A clean ending to the opening act. Two years of building a startup while studying, shipping real systems, and learning what it means to own something end-to-end. Ready for the harder problems.
Bachelor's degreeNortheastern · Going Deeper
Started my MS in Computer Science at Northeastern University. Going deeper on AI systems — RAG pipelines, semantic search, LLM applications in production. Also building SqlSensei and Workday Jarvis. The work doesn't stop between semesters.
MS CS · NortheasternStill writing
About
I started with a C compiler — not a bootcamp.
By 2023 I co-founded a startup, built a remote code execution engine from scratch, and ran it in production for two years. In 2024 I was building computer vision pipelines for the Government of India. In 2025 I'm at Northeastern, going deeper on AI systems. I haven't stopped shipping.
Running a startup
Production uptime
Daily users monitored
APIs shipped at IIIT-H
What I bring
Founder Energy
Co-founded Dcrypt and ran it in production for 2 years — microservices, Docker, RabbitMQ, 99.9% uptime. I know what it means to own a system end-to-end.
Systems Thinker
Whether it's a code execution engine or a CV inference pipeline over 200 camera feeds — I think in systems, not scripts. Architecture decisions matter.
AI-Native
RAG pipelines, semantic search, private inference with Ollama — built into real products, not proof-of-concept notebooks. LLMs are an engineering problem.
Full-Stack Depth
React to RabbitMQ. Postgres to Nginx. TypeScript to Python. I can go from designing a data model to shipping a pixel-perfect UI in the same sprint.
The longer version
My path didn't start at a bootcamp. It started with a C compiler, a lot of manuals, and problems I didn't know how to solve. By 2021 I was building Java games — Hunt the Wumpus — and realizing that software felt like worlds you could design.
In 2022 I was freelancing. Real clients, real deadlines, real consequences for late or broken code. A year later I co-founded Dcrypt, a coding education startup powered by a custom remote code execution engine — Docker sandboxed containers, async job queues on RabbitMQ, Redis for caching. I ran it for two years and maintained 99.9% uptime. That taught me what it means to own infrastructure end-to-end.
MS in Computer Science
Northeastern University · Jan 2025 – Present
Selected Work
Things I've built
From a startup I co-founded to AI systems in production — each one was a real problem, not a portfolio piece.
CollaBridge
A platform connecting brands with influencers — campaign to contract.
Full-stack collaboration marketplace where brands post campaigns, influencers apply, and both sides manage the entire workflow — applications, approvals, and active collaborations — in one place.
Two-sided marketplace with role-separated brand and influencer dashboards
Full campaign lifecycle: post → apply → approve → collaborate
JWT auth with role-aware protected routing
SocialSynth
Turn one idea into platform-ready content for every channel.
AI content repurposing tool — paste a blog post, transcript, or idea and get tailored captions, threads, and posts generated for LinkedIn, Twitter, and Instagram in seconds.
SqlSensei
An AI SQL tutor with live query execution and real-time feedback.
LLM-powered SQL learning platform — translate natural language into executable SQL, run it against a live MySQL backend, and get contextual explanations and error feedback in real time.
SOP Generator
Describe a process in plain English — get a production-ready SOP.
AI tool that converts plain-language process descriptions into structured, professional Standard Operating Procedures with numbered steps, roles, decision points, and formatting.
Recipe Roulette
Spin for a recipe. Tell it what's in your fridge. Never eat boring food again.
AI-powered recipe discovery app — input what ingredients you have and get a randomized, cookable recipe with instructions, substitutions, and a shopping list for what's missing.
Career
Experience
Real ownership, real systems, real load. Not just code review comments.
DCRYPT
Founding Software Engineer
Architected and ran a full-stack coding education platform from zero — microservices, Docker-based code execution, and async job pipelines — in production for two years.
- Architected a microservices-based platform capable of supporting thousands of concurrent users — horizontally scalable from day one.
- Designed and deployed a containerized remote code execution engine using Docker, securely running Python, JavaScript, Java, C, and C++ workloads; reduced average execution latency by 50%.
- Implemented async job processing pipelines using RabbitMQ, decoupling execution services and increasing system throughput by 40%.
- Integrated Redis for distributed caching and session management, reducing API response latency by 35% under peak load.
- Built multi-environment infrastructure (dev, staging, production) with automated deployment workflows, cutting deployment-related failures by 70%.
- Deployed on Ubuntu-based cloud infrastructure with reverse proxies, process monitoring, and observability — maintained 99.9% service availability across 2 years.
IIIT-H
Software Research Intern
Built real-time crowd monitoring infrastructure and CV inference pipelines for the Government of India's Smart City Lab — serving 10K+ daily commuters across 20+ monitored zones.
- Contributed to a real-time crowd monitoring system deployed under the Smart City Lab (Government of India), supporting operational monitoring for 10K+ daily commuters.
- Designed and implemented 15+ RESTful APIs for analytics dashboards, real-time people-count tracking, and automated crowd-density alerting across 20+ monitored zones.
- Integrated YOLOv5-based computer vision models into the backend inference pipeline using Dockerized services — enabling 90%+ detection accuracy with sub-200ms inference latency.
- Built and optimized video ingestion and processing pipelines supporting 200+ concurrent camera feeds under continuous operational load.
- Automated ML model versioning and deployment workflows using scheduled inference jobs, enabling zero-downtime model updates and reducing deployment turnaround by 35%.
MS Computer Science
Northeastern University · Boston, MA
Capabilities
What I work with
Organized by capability, not by logo. Each area reflects real production use.
Product Engineering
Building interfaces and experiences that feel fast, intentional, and complete — from data model to pixel.
AI Systems & LLMs
Designing pipelines that actually work in production — RAG architectures, semantic search, private inference.
Backend Architecture
Systems that handle real load — microservices, async job queues, caching layers, and the infrastructure that runs them.
Applied ML & Computer Vision
Translating research into working inference pipelines — from training to deployment at operational scale.
Data & Databases
Relational, document, vector, and cache stores — picking the right one and using it correctly.
Cloud & Infrastructure
Deploying systems that stay up — multi-environment pipelines, containerization, observability.
Right Now
What I'm focused on
Current chapter. Active interests. What's taking up most of my headspace in 2025.
AI Systems & RAG
Building retrieval-augmented generation pipelines that work in production, not just in demos.
Developer Tools
Tooling that reduces friction for engineers — fast, composable, and opinionated in the right places.
Scalable Backends
Microservices, async job queues, and distributed systems that stay reliable under real load.
MS Research @ Northeastern
Graduate coursework and research pushing into the harder problems in computer science.
Product Engineering
Thinking about what should be built — not just how to build it — and closing that gap fast.
Startup Experiments
Testing new ideas in the space between AI, developer experience, and useful software.
Let's Connect
Get in touch
Open to full-time engineering roles, research collaborations, and projects with real ambition. If you're building something hard, let's talk.
Prefer a direct line? Find me on any of these:
GitHub
@abdullahbasarvi
Abdullah Basarvi
abdullahbasarvi@gmail.com
Typically respond within 24 hours.