AI Engineer / Bhopal / 未来を作る
I turn AI, Android, and web experiments into things people can actually use.
Fresher and 2027 graduate based in Bhopal. Search my full name online and you will find the same pattern: projects, experiments, certificates, hackathons, and a steady habit of building in public.
Nine projects I keep returning to: personal AI, local systems, Android utilities, document intelligence, semantic search, computer vision, and practical product tools.
See Works
AiOS Assistant
(01)
Local-first personal AI infrastructure with a dashboard, browser connectors, desktop workers, and explicit action controls.
What Do You Do
(02)
Privacy-first digital wellbeing dashboard with local desktop activity tracking, clear review controls, and a calmer relationship with screen time.
Kaamurai OS
(03)
Agent-first Arch Linux: describe an outcome, review the plan, then let a local system agent carry out the work safely.
LegalEase
(04)
Legal document intelligence with summaries, conversational assistance, editable outputs, and an open-source contributor path.
Advanced Optimizer
(05)
Android battery intelligence that makes app and sensor drain visible before proposing an on-device optimization plan.
GlassInterface
(06)
Offline assistive vision with object recognition, distance estimation, risk scoring, tracking, voice commands, and spoken guidance.
Cognix
(07)
Semantic search with MiniLM embeddings, ChromaDB retrieval, probabilistic clustering, and a cache that learns repeated intent.
FlightIQ Ultimate
(08)
Semantic destination discovery, vector search, route optimization, recommendations, analytics, and natural-language SQL over flight data.
Fasal Sathi
(09)
Hindi-first Android support for weather, crop guidance, voice and camera help, recommendations, and government schemes.
Nine focused builds shaped for real briefs, fast demos, and public proof.
01 / Evidence AI
Evidence-linked timelines, entity resolution, graph analysis, next-step support, and review-ready reports.
02 / ET AI 2026
Traces changes across operational documents, surfaces consequences, and keeps approval with the human operator.
03 / OpenAI Build Week
Turns public GitHub work into an evidence-backed learning profile and a four-week build sprint.
04 / GenAI APAC
A smart-city decision workspace that organizes civic signals while keeping final actions human approved.
05 / India.runs
Recruiter-grade candidate ranking with transparent evidence, risk signals, and an offline ranking path.
06 / Built for Good
Jobs, income support, and a portable Skill Passport designed around informal workers.
07 / Meta AI
An OpenEnv-compatible environment where agents clean messy DataFrames through deterministic actions and measurable rewards.
08 / HackVortex
Satellite-image segmentation and web mapping for surfacing likely sewage infrastructure patterns.
09 / FOSS Hack 2025
A preserved open-source hackathon repository. Kept here as participation proof without inventing a product story the repo does not claim.
Challenge work that became complete, testable interfaces rather than a deck.
World Cup 2026 operations and fan intelligence with live command views, Gemini-assisted workflows, and deterministic fallbacks.
A non-partisan guide to Indian election processes with useful Google service integrations and resilient fallback flows.
As a GSSoC'26 Project Admin for LegalEase, I am shaping contributor-friendly issues, clearer AI document flows, and a healthier path from first contribution to useful release.
(6)
I like turning a chat box into a real workflow with memory, planning, retrieval, and a control surface people can understand.
Portfolios, product pages, control rooms, and web apps that feel considered instead of looking like a template.
I enjoy practical mobile ideas: assistive vision, quiet utilities, and AI helpers that make sense away from a laptop.
I treat prompts as interaction design: shaping context, guardrails, and the little decisions that make an AI flow feel useful.
Data, documents, AI outputs, and operations become more interesting when the interface helps you see what matters quickly.
Give me a messy problem and a timer. I enjoy narrowing the scope, making a strong demo, and shipping something people can try.
Culture, sound, interfaces, and stories that keep the build energy moving.
Real music fan here: K-pop, J-pop, late-night loops. Let's connect on Spotify <3
Rosé
My favorite music artist and singer: voice, personality, and yes, my bias.
Voice / Bias
K-pop
Mostly TWICE, LE SSERAFIM, ILLIT, NewJeans, and whatever loops all week.
Playlist fuel
AI
IYKYK: the real AI lol. Agents, interfaces, model behavior, weird systems.
IYKYK
Anime
Black Clover, MHA, SAO, Dr. Stone. Ranking is hard; I like too many.
Worldbuilding
Gaming
Mostly Valorant nowadays: pressure, comms, timing, tiny improvement loops.
Valo phase
日本
Learning Japanese, loving J-pop, and yes: recruit me Japan TwT.
J-pop / LanguageLearning, hackathons, open source, and systems that became real enough to show.
Started formal CS learning, built the Python and web foundations, and began collecting the small proofs that made later projects possible.
Moved from tutorials to public experiments: web interfaces, Python utilities, CTF participation, and the first hackathon builds.
Turned AI ideas into product flows across Android, document intelligence, semantic search, and user-facing dashboards.
Expanded from individual apps into privacy-first tools, agentic systems, investigation intelligence, and open-source product work.
I am not trying to look experienced. I am trying to make the proof impossible to ignore: projects, certifications, timelines, and an interface that remembers the work.
