I have worked on a bunch of projects for my college peers and seniors, usually starting from a rough idea and turning it into a complete working product. I am comfortable handling backend logic and the smart parts of a project, and I vibe code the entire build end to end. I adapt quickly and focus on shipping something solid rather than overthinking every detail. If you need someone who can take an idea and just make it real, I can do that.…

Archit Rawat

I have worked on a bunch of projects for my college peers and seniors, usually starting from a rough idea and turning it into a complete working product. I am comfortable handling backend logic and the smart parts of a project, and I vibe code the entire build end to end. I adapt quickly and focus on shipping something solid rather than overthinking every detail. If you need someone who can take an idea and just make it real, I can do that.…

Available to hire

I have worked on a bunch of projects for my college peers and seniors, usually starting from a rough idea and turning it into a complete working product. I am comfortable handling backend logic and the smart parts of a project, and I vibe code the entire build end to end. I adapt quickly and focus on shipping something solid rather than overthinking every detail. If you need someone who can take an idea and just make it real, I can do that.

See more

Skills

Experience Level

Python
Expert
SQL
Expert
TensorFlow
Expert
Java
Expert
Keras
Expert
MongoDB
Expert
MySQL
Expert
Docker
Expert
GitHub CoPilot
Intermediate
AWS
Intermediate
Google Cloud
Intermediate
SQLite
Intermediate
PyTorch
Intermediate
Kubernetes
Intermediate
See more

Work Experience

Software Engineer (LLM Inference Platform)
January 1, 2026 - Present
Engineered a production-grade multi-model LLM inference platform with semantic request routing. Implemented load-aware request optimization and caching to improve stability and latency under concurrent traffic. Load-tested using Locust, analyzing RPS, error rates, and p50-p99 latency to identify scalability bottlenecks.
Software Engineer (Disaster Response Platform) at Alertrix
January 1, 2025 - Present
Built a modular FastAPI backend enabling AI-driven disaster response classification, summarization, geolocation, and severity scoring. Created a real-time reporting dashboard and map-based alert interface using FireBase for live incident visibility.
Software Engineer (NLP / ML) - Real-Time Hate Speech Detection
January 1, 2025 - Present
Built a real-time NLP inference service using a fine-tuned Transformer model for hate-speech detection. Improved dataset balance with augmentation, achieving 94% F1 score on diverse text samples. Packaged as a Docker container and deployed on Hugging Face Spaces, achieving sub-300 ms average inference latency.

Education

Bachelor of Technology in Computer Science Engineering at Lovely Professional University, Phagwara, Punjab
September 1, 2022 - July 1, 2026
Senior Secondary Certificate (12th) at The Horizon School, Dehradun, Uttarakhand
April 1, 2019 - March 1, 2021
Secondary School Certificate (10th) at Doon Public School, Dehradun, Uttarakhand
April 1, 2018 - March 1, 2019

Qualifications

Applied Generative AI Certification – InfoSys Springboard
May 1, 2025 - February 7, 2026
Cloud Computing – NPTEL / SWAYAM
October 1, 2024 - February 7, 2026
Data Structures and Algorithms – Board Infinity
July 1, 2024 - February 7, 2026

Industry Experience

Software & Internet, Education, Professional Services
    paper Multi-Model LLM Inference API

    Designed and built a production-grade, multi-model LLM inference platform capable of serving concurrent workloads at scale. The system supports semantic request routing to dynamically select the most appropriate model per request, improving response quality and efficiency. To ensure reliability under load, the platform incorporates load-aware request optimization and intelligent caching, significantly reducing latency and improving stability during traffic spikes. Scalability and performance were validated through extensive load testing with Locust, analyzing requests per second, error rates, and p50–p99 latency metrics to identify and address system bottlenecks. Currently working to add caching and observeability