Samir Kharel
Machine Learning Engineer & Full Stack Developer
Building production AI systems — computer vision, multi-agent RAG pipelines, and applied ML — backed by multiple national hackathon wins and a Quantum Computing diploma.
- Jhapa, Nepal
- Project Manager, Mavtech
- Mechi Multiple Campus, Bhadrapur — Currently 8th semester
- 13
- 3
- 98%
- 3+
Selected work
Four systems that shipped, with the problem, what I built, and where each ended up.
Road Sense
Real-time road accident detection & alert system
In Nepal, delayed emergency response to road accidents significantly increases casualties. No automated, affordable detection system existed for rural roads.
Built the full detection pipeline: frame ingestion, YOLO-based accident classification, and the real-time alert dispatch system.
2nd Runner-up at Codefest National hackathon
Python · OpenCV · YOLO · TensorFlow · IoT · Raspberry Pi
AI Talent Intelligence Platform
Multi-agent candidate screening & recruiter copilot
Recruiters spend hours manually screening CVs. Candidates lack specific, actionable feedback on skill gaps for a given role.
Designed and built the multi-agent orchestration layer, the RAG pipeline for semantic resume parsing using Qdrant vector search, the skill gap radar chart module, and the interview simulation agent.
End-to-end semantic parsing with dynamic interview generation
Python · FastAPI · Next.js · Qdrant · LangChain · OpenAI · RAG · Docker
Nigrani
Elephant detection system that protects village communities
Human-wildlife conflict in Nepal's border communities costs lives and livelihoods. Existing solutions (rangers, fences) are expensive and slow to respond.
Built the detection model (YOLO trained on custom dataset), deployed inference to Raspberry Pi edge nodes, and implemented the SMS/siren alert pipeline.
1st Runner-up at BPC Hackfest National — deployed concept validated
Python · YOLO · OpenCV · Raspberry Pi · IoT · TensorFlow Lite · SMS API
WhatsApp Automation
LLM-powered business automation via WhatsApp
Small businesses in Nepal lack affordable 24/7 customer support. WhatsApp is the dominant communication channel, but no smart automation existed for it.
Built the LangChain agent pipeline, integrated Supabase pgvector for RAG-based knowledge search, implemented dynamic prompt management, and wired the Meta WhatsApp Cloud API webhook flow.
Live deployment handling real customer traffic
LangChain · Supabase · pgvector · Next.js · Python · Meta Cloud API · Gemini · OpenAI
Technical Skills
Nothing here is a bar chart pretending to be data. Skills sit in the domain they belong to, and an edge means I have actually used the two together in a shipped project.
Hover or tab a node to light up everything it is entangled with.
ML / AI
Backend
Frontend
Quantum
Tools & Infra
Roles and community
Client software delivery, and the developer community I build around it in Koshi Province.
Project Manager
Mavtech
- Led technical development and production deployment of multiple full-stack web and mobile platforms
- Owned feature planning, system architecture design, and release workflows for client software products
- Coordinated cross-functional teams to deliver scalable solutions on-schedule for corporate clients
Full Stack Developer
Kisoft Tech Solutions
- Built custom, responsive web applications for corporate and retail clients to business specifications
- Implemented clean frontend interfaces and RESTful backend services with seamless cross-browser support
Community & leadership
President
BCA Association, Mechi Multiple Campus
- Organized GitHub Copilot Dev Days Jhapa — flagship regional developer conference
- Directed Student Empowerment Program: hands-on technical training and career mentorship
- Led community to multiple national hackathon placements
Community & Event Lead
Mechi Mavericks
- Organized CodeWar and BCA Roadmap developer workshops across Koshi Province
- Mentored software developer teams; led teams to provincial and national tournament victories
Competitions and credentials
Results stated exactly as they were placed, with the certifications and academic record behind them.
Competitions
1st Runner-up
Elephant detection & alert system using ML and IoT. Deployed along village perimeters to reduce human-wildlife conflict.
2nd Runner-up
Road accident detection and real-time alert system using machine learning.
Winner
1st place, organized by Codefest. Team: Mechi Mavericks.
Participant
International space hackathon, hosted by NRCC Nepal.
Participant
Quantum computing hackathon hosted at MIT. One of the toughest quantum competitions globally.
Certifications
Quantum Computing & Programming
OpenCV Bootcamp
Generative AI
Education
Bachelor in Computer Applications (BCA)
+2 (High School in Science)
How I work
Background
I'm a machine learning engineer based in Jhapa, Nepal, with hands-on experience across computer vision, deep learning, RAG pipelines, and multi-agent systems. My work spans real-world deployments — from CCTV-based elephant alert systems along village edges, to LLM-powered WhatsApp automation handling live customer traffic. I'm drawn to problems where the gap between a research idea and a working system is still wide, and I want to be the engineer who closes it. I hold a Quantum Computing diploma from QNepal (2023), participated in MIT QHack, and work with Qiskit — the theme of this site is earned, not decorative.
Quantum Computing diploma · QNepal 2023 · MIT QHack participant · Qiskit practitioner
Get in touch
Open to machine learning engineering roles. Email reaches me fastest.