PROJECT / 01
AI Agent Workflow Automation Platform
An LLM-powered agent platform orchestrating multi-step tasks through structured prompt pipelines and a YAML-driven config layer for composing new workflows without code changes.
STATUS: OPEN_TO_OPPORTUNITIES
Eklakh Dewan
AI Engineer / Data Scientist
Building AI that explains itself.
AI Engineer & Data Science undergraduate at KPRIET, shipping end-to-end ML systems — from FastAPI inference services to agentic LLM workflows — with explainability and reliability treated as first-class requirements, not afterthoughts.
NLPCore Focus
6+Shipped Projects
3Certifications Live
2027Graduating
inference_trace.json
01 INPUT
A B.Tech engineer who treats ML like software, not a notebook experiment.
Specializing in NLP, agentic LLM workflows, and explainable ML — with hands-on production experience across data pipelines, REST APIs, and deployment. Full profile and education on the About page →
02 PROCESS
Core stack
The tools this work actually runs on — confidence reflects depth of hands-on production use, not just familiarity.
Languages
AI / ML
Engineering
03 OUTPUT
Featured work
Six shipped projects spanning agentic workflows, explainable recommenders, and multilingual NLP. Full case studies on the Projects page.
PROJECT / 02
AI Career Path Recommender
Explainable ML recommendation engine using SHAP to surface why each suggestion was made, served through a real-time FastAPI backend.
PROJECT / 03
Job Recommendation Engine
Resume-to-job semantic matching using BERT embeddings, TF-IDF, and cosine similarity — with a skills gap-analysis module recommending learning paths.
04 DEPLOY
Let's ship something that works in production, not just in a notebook.
Open to AI/ML internships, research collaborations, and software engineering roles. Reach out — I reply fast.