Egocentric Video Processing & Quality Control Platform
Separating useful work footage from noise in hours-long first-person video — with pose-level QC.
> role:Generative AI Engineer, Forward Deployed AI Engineer, AI & Computer Vision Engineer, Full-Stack Developer, IoT & Robotics Builder
I build machines that see, think, sense and move.
I build AI-powered applications, computer vision pipelines, intelligent automation systems, and scalable full-stack products.
Select system · retunes the site
Servos · ESP32 · embodied AI
01
I'm an engineer interested in building intelligent systems that connect AI research with practical applications.
My work spans Generative AI, LLM applications, computer vision, full-stack development, IoT, and robotics.
I enjoy solving engineering problems end-to-end — from understanding requirements and designing architectures to developing applications, integrating models, testing on real-world data, and deploying working solutions.
I've worked on AI-based video analytics, object detection and tracking, intelligent automation, voice-interactive robotics, agricultural monitoring, and distributed applications.
The best technology isn't just technically impressive — it solves a genuine problem and works reliably in the real world.
Generative AI
LLM apps, RAG, agents, evaluation
Computer Vision
Detection, tracking, pose, video
Full-Stack
React, Spring Cloud, APIs
IoT & Robotics
ESP32, LoRa, MQTT, servos
B.Tech, Internet of Things
KL University · 2022 – 2026

The journey so far
2022
Started B.Tech in IoT
KL University — began building with ESP32, sensors and computer vision.
2024
Hardware + vision builds
Face-tracking device, ball- and human-following robots with OpenCV and ESP32.
Aug 2024
Web3 Security Data Analyst
Webacy.
Oct 2024
LLM Evaluation & Prompt Engineering
Outlier AI — evaluating model outputs in human-in-the-loop workflows.
Nov 2024
Blockchain & Full-Stack Intern
Sky Trade.
2025
AI data & evaluation
Mindrift — annotation and model response evaluation. Built RAG agents and AI robotics projects.
Apr 2026
Full Stack Engineer Intern
Cognizant, Chennai — Spring Cloud microservices and React.
Jun 2026
Claude Certified Architect
Anthropic — Foundations.
Now
Video intelligence platform
Building a VLM- and pose-driven video QC pipeline.
02
From Spring Cloud microservices at Cognizant to evaluating LLM outputs in human-in-the-loop workflows. Each role is listed with its real dates.
Built ReleasePilot — a Jira-inspired platform for managing products, roadmaps, backlogs, releases, environments and QA gates — on a React/Vite frontend and a Spring Cloud microservices backend.
03
One case study per discipline — generative AI, computer vision, robotics, IoT and full-stack — each built around a concrete problem.
Separating useful work footage from noise in hours-long first-person video — with pose-level QC.
Live pipeline · egocentric video QC
Stage 1 / 11
Upload
Long-duration videos are uploaded and registered with their metadata.
Target architecture of a project in active development — not every stage is complete.
Turning raw CCTV footage into structured, reviewable vehicle events.
03A 3D-printed humanoid that sees, talks and shakes hands.
04Distributed sensor nodes watching soil, air and climate in real time.
A Jira-inspired product lifecycle platform on Spring Cloud microservices.
04
How I approach a real-world problem: start with the people and constraints, design the whole system, then build, measure and iterate. Click any step.
Step 1
Understand the problem
05
Grouped by domain, with an honest level for each: experienced, hands-on, or familiar.
Prompting, evaluation, retrieval and agents.
Detection, tracking, pose and classical CV.
Web apps, APIs and microservices.
What I write code in.
Embedded systems that touch the physical world.
Storage, infrastructure and delivery.
06
4×
Hackathon winner
4×
Newspaper & TV features
9.34
CGPA, B.Tech IoT
40+
Public repositories
Hackathons & community
IIT BHU Eco Hackathon
Recognition at the IIT BHU Eco Hackathon.
Unplugged 2.0
Hackathon achievement.
HackWithIoT
Helped organize and host a 24-hour college hackathon.
Hardware, AI & IoT hackathons
Regular participant across innovation and hardware events.
In the media
Humanoid robotics
Luna, the AI-powered humanoid robot.
LoRa gateway & smart street lighting
Long-range IoT for city infrastructure.
AI / IoT innovation
Applied AI and IoT student projects.
Certifications
Claude Certified Architect — Foundations
Anthropic · June 2026
AI Fluency: Framework & Foundations
Anthropic
Oracle Cloud Infrastructure Foundations Associate
Oracle
07
08
Engineering notes are on the way.
I'll be writing about what I learn building real systems — the parts that break, and what fixes them.
Planned topics
09
I'm always interested in connecting with engineers, founders, researchers, and teams building meaningful AI-powered products.
@icrohan
Live from the GitHub API
Primary languages across repositories