FactoryBrain AI
A local-first manufacturing knowledge prototype for organizing, searching, and exporting operational documents.
Python · Streamlit · Local-First Architecture

APPLIED AI / AI CONSULTING
Mechanical engineering background, now focused on practical AI workflows, evaluation, implementation planning, and AI-assisted prototyping.
ABOUT ME
From everyday curiosity to exploring how AI can solve real problems.
I first discovered ChatGPT through an article about a new kind of tool that could answer questions on almost anything. I followed the link and started experimenting out of curiosity. Before long, AI had become part of my everyday life, and my curiosity kept taking me further.
I began with learning, then moved into research, writing, idea development, and creative work. As I explored more, I started using different AI models and tools for different purposes, discovering their strengths, limitations, and the kinds of tasks they handled best. I became increasingly interested not just in what AI could do, but in how the right context, instructions, and iteration could turn an ordinary response into something genuinely useful.
Eventually, I moved from using AI to building with it. With AI-assisted coding tools, I developed FactoryBrain AI, a manufacturing knowledge prototype, and a crypto/VDA reconciliation case study. Along the way, I learned to refine instructions, question outputs, and improve results rather than simply accept the first response. For me, working effectively with AI means understanding the task, choosing an appropriate approach, evaluating the result, and knowing when further refinement or human judgment is needed.
My manufacturing experience gave me a practical foundation for this direction. At Jindal Stainless, I worked with production reports, SAP, Excel, and operational information, seeing how much everyday work depends on accurate data, clear reporting, and reliable processes. I'm now building on that experience by learning AI automation and exploring how businesses can identify bottlenecks, improve repetitive workflows, and determine where AI can genuinely help, rather than adding technology for its own sake.
My long-term ambition is to bridge the gap between business problems and practical AI implementation. I want to help businesses identify meaningful opportunities, design useful workflows, and turn promising ideas into working solutions. Through independent projects, ongoing learning, and paid AI model evaluation work, I'm developing the practical experience to move toward AI consulting and implementation.
I also explore generative AI as a creative medium, experimenting with posters, image generation, and visual concepts. I enjoy the challenge of turning an idea into something visually compelling, refining the results through experimentation and creative judgment.
Outside my professional interests, I enjoy following sports, particularly football, and reading detective fiction.
EDUCATION
Diploma in Mechanical Engineering — Purulia Polytechnic, 2025
EXPERIENCE
Trainee Process Control Engineer — Jindal Stainless Limited, Odisha
SELECTED WORK
A local-first manufacturing knowledge prototype for organizing, searching, and exporting operational documents.
Python · Streamlit · Local-First Architecture

A human-in-the-loop workflow for reconciling digital-asset transaction data across multiple exchanges — emphasizing traceability, exception handling, and human verification.
Multi-Source Reconciliation · Privacy-Safe · Human-Centered

PRACTICAL AI & WORKFLOW CAPABILITIES
I help businesses understand challenges, identify relevant AI opportunities, research tools, design workflows, and develop small prototypes or implementation plans.
Identifying where AI can practically help — and where it can't. Researching tools, analysing use cases, and developing focused strategies grounded in real business needs.
Designing practical workflows that integrate AI into existing processes. I also consider human oversight and verification when designing AI-assisted workflows.
Building lightweight working prototypes to test ideas before investing in larger systems. Planning implementation approaches that match the actual scope of the problem.
Evaluating AI tools, models, and outputs for practical usefulness. Testing whether things actually work, identifying limitations, and providing honest assessments.
Organising, cleaning, and analysing data. Building reporting workflows, dashboards, and operational documentation using structured data processes.
Developing content using structured prompting and AI-assisted creative workflows. Supporting short-form writing, visual storytelling, and professional communication.
TOOLKIT & TECHNOLOGY STACK
I work across a growing ecosystem of AI models, research platforms, development environments and creative tools. I use them to explore ideas, solve practical problems, analyse information and build AI-assisted solutions.
My focus is not simply on using different tools, but on understanding where each one can add value.
My primary AI workspace for planning, reasoning, prompt development and iterative problem-solving.
Used for in-depth research, document analysis and exploring complex topics.
An alternative AI assistant for critical reasoning, writing, complex tasks and code review.
Used for everyday AI assistance, information exploration and productivity.
Exploring alternative AI models and their reasoning capabilities.
Experimenting with alternative AI models and their capabilities.
Source-grounded analysis of PDFs, YouTube videos and research materials, including brainstorming and report preparation.
Research, information discovery and finding relevant sources.
Exploring discussions, emerging trends and conversations on X, particularly around Web3.
AI-assisted development and iteration of portfolio projects and practical prototypes.
AI-assisted coding, implementation and refinement of project workflows.
Exploring visual workflow automation through triggers, modules, data mapping and integrations.
Experimenting with generative workflows, existing templates and static ad-builder concepts.
AI-powered image generation and visual experimentation.
Experimenting with AI-powered video generation.
AI-powered image generation and editing.
Exploring AI video generation through hands-on experimentation with its free version.
Voice generation and voice cloning.
Voice-to-text conversion and transcription.
HOW I WORK
Every project follows the same practical framework — understand the problem, reduce it to its simplest form, build something testable, and document what works and what doesn't.
Define the real problem. Research the context, identify where AI may genuinely help, and determine what a useful outcome looks like.
Reduce the problem to its smallest useful version. Cut unnecessary complexity and focus on what needs to be solved first.
Build a lightweight working version to test the idea. Use the simplest reasonable tools and architecture before investing in anything larger.
Check whether the prototype actually works. Identify what breaks, what’s missing, and where the limitations are.
Record decisions, workflows, results, and limitations. Make it possible for someone else to understand and continue the work.
GENERATIVE AI / VISUAL WORK
A collection of visual experiments exploring image generation, visual direction, composition, typography, and creative possibilities.

Rain, plants, thoughts, repeat — ambient monsoon telemetry.
MODEL USED: CHATGPT IMAGE 2.5

Leeds underground late licence club night in neon acid lime motion silhouettes.
MODEL USED: CHATGPT IMAGE 2.5

Warm roasted notes and sweeping calligraphic vermilion script.
MODEL USED: CHATGPT IMAGE 2.5

Electric Klein blue framing a decagram starburst rave crowd aperture.
MODEL USED: CHATGPT IMAGE 2.5

Entomological taxonomy in neon fluorescent ink and station telemetry.
MODEL USED: CHATGPT IMAGE 2.5

Echoing condensed typography revealing an indigo-to-amber sunset horizon.
MODEL USED: CHATGPT IMAGE 2.5

Technical hooded shell garment release with strikeout inventory checklist.
MODEL USED: CHATGPT IMAGE 2.5

Triple motion-blurred runner silhouettes cutting through pre-dawn mist.
MODEL USED: CHATGPT IMAGE 2.5

Old paintings, new code — classical riverside landscape with glowing terminal statue.
MODEL USED: CHATGPT IMAGE 2.5

Carved smiley relief framed by ornate lotus florals and raw kraft borders.
MODEL USED: CHATGPT IMAGE 2.5

Lo-fi rubber-stamped basement show flyer with repeating halftone bubbles.
MODEL USED: CHATGPT IMAGE 2.5

Fresh • Flaky • Forever — giant woodcut croissants suspended in azure cloud cover.
MODEL USED: CHATGPT IMAGE 2.5

Sunglasses-wearing llama at an overgrown CRT desk amidst corporate chaos.
MODEL USED: CHATGPT IMAGE 2.5

Split-second punch impact in crimson duotone halftone and cobalt grease marks.
MODEL USED: CHATGPT IMAGE 2.5
CERTIFICATIONS & LEARNING
Targeted coursework and professional specializations validating structured prompt design, workplace AI workflows, and responsible implementation.

Google Coursera
Four-course series covering structured prompt design, task framing, multi-turn AI collaboration, data analysis, and presentation workflows using generative AI tools.

Google Coursera
Five-course foundational program covering practical generative AI usage, workplace productivity tools, prompt creation basics, and responsible, ethical AI practices across daily workflows.

Microsoft and LinkedIn LinkedIn Learning
Learning path covering core generative AI concepts, Microsoft Copilot practical usage, and responsible AI principles for workplace problem-solving and task support.