Available for opportunities

Debasmita Chatterjee

AI/ML Engineer specialising in LLM systems, Generative AI pipelines, and production-grade machine learning — turning research into real deployable systems.

0
LLM responses evaluated
0
ECE reduction achieved
0
Projects deployed live
Scroll
01 — About me

Building AI that actually works.

DC
B.Tech CSE · LPU Punjab

I'm a final-year Computer Science student at Lovely Professional University with a Minor in Data Science, passionate about building AI systems that go beyond prototypes and into production.

My work spans LLM evaluation, calibration research, Generative AI pipelines, and end-to-end ML deployments. I've evaluated 1600+ LLM responses professionally at Outlier.ai and built multiple live AI systems from scratch.

I'm drawn to the gap between "it works in a notebook" and "it works in the real world" — and I build things that bridge it.

02 — Technical skills

The stack I build with.

From raw model inference to full-stack deployment — across the entire ML lifecycle.

🧠
ML / AI
PyTorch TensorFlow Scikit-learn Transformers LangChain LangGraph LlamaIndex FAISS RAG
⚙️
Frameworks & Tools
FastAPI Flask Streamlit React Node.js Docker GCP REST APIs Git
💻
Languages
Python SQL JavaScript Java C/C++ R HTML/CSS
📊
Data & Analytics
Pandas NumPy Matplotlib Feature Engineering Model Validation Information Retrieval Neural Networks
03 — Experience

Where I've worked.

Outlier.ai Freelance · Remote
AI Trainer — LLM Evaluation
Aug 2025 — Jan 2026
  • Evaluated 1600+ LLM responses using structured rubrics, identifying hallucinations, reasoning errors, and failure modes to support RLHF pipelines and model evaluation workflows.
  • Conducted comparative ranking and preference annotation across reasoning, multilingual, and domain-specific tasks to improve model response quality.
  • Applied standardized evaluation frameworks and taxonomies to ensure consistency and reliability in large-scale annotation workflows.
  • Generated actionable feedback for model improvement, contributing to enhanced accuracy, coherence, and alignment of LLM outputs.
Lovely Professional University Education · Ongoing
B.Tech in Computer Science — Minor in Data Science
2022 — 2026
  • Specialising in ML, AI, and data science with a strong foundation in algorithms, statistics, and systems design.
  • Self-driven project work in LLM calibration, Generative AI, and production ML systems beyond the curriculum.
04 — Projects

Things I've built.

Production-deployed AI systems — not just notebooks.

🎯
LLM Confidence Calibration & Overconfidence Analysis
Production-grade statistical framework diagnosing and correcting overconfidence in Mistral-7B-Instruct and Phi-2 via logit-level ECE measurement and post-hoc temperature scaling — without any retraining.
62% ECE reduction 81.2% accuracy preserved 1000-iter bootstrap CI
PyTorch HuggingFace Mistral-7B Phi-2 Streamlit SciPy
🚀
AI Agency Workflow Automation Platform
End-to-end ML pipeline automating lead scoring, classification, and proposal generation for AI agencies. From client form submission to tailored proposal in under 5 seconds with LangGraph orchestration and RAG-powered context.
<50ms ML inference 6-stage pipeline Live deployed
FastAPI LangGraph FAISS RAG Whisper Streamlit
💹
RL-Based Dynamic Pricing Engine
Proximal Policy Optimisation (PPO) agent trained to maximise profit in dynamic market conditions — simulating demand elasticity, competitor pricing, and inventory constraints with a custom Gymnasium environment.
PPO Actor-Critic 200 episodes Live Streamlit app
TensorFlow Gymnasium PPO NumPy Streamlit Plotly
05 — Achievements & Certifications

Recognised work.

🏆
ET GenAI Hackathon — Round 2 Selection Selected from 54,618+ registrants. Built and submitted a functional end-to-end AI system prototype demonstrating production-ready design.
☁️
Oracle Cloud Infrastructure 2025 — Generative AI Professional Validated skills in LLMs, RAG, vector search, LangChain, and OCI GenAI services for building production-ready AI applications. Oct 2025.
🤖
Oracle Cloud Infrastructure 2025 — AI Foundations Associate Demonstrated strong understanding of AI/ML fundamentals, deep learning, LLMs, and OCI AI services. Mar 2026.
📝
LLM Evaluation at Scale — Outlier.ai 1600+ professional LLM response evaluations supporting RLHF pipelines across reasoning, multilingual, and domain-specific tasks.
06 — Contact
Let's build something together.

I'm actively looking for ML/AI engineering roles and internships.
Reach out — I reply fast.