App LogoApp name
BB

Bhanupratap Bhadouria

Open to work2 years experience

AI Engineer

Pune, IndiaTarget Roles: Backend Engineer • Full-Stack Developer • Frontend Specialist

AI Engineer specialising in LLM integration, RAG pipelines, and end-to-end ML systems.

GitHub

Standing Rank

Rank Not Available

Developer Badges

No badges earned yet

Skills & Technologies

53 skills
rag pipelinesllm integrationnlptransformersembeddingssemantic searchshap explainabilitylangchainlanggraphhuggingfacepytorchtensorflowscikit-learnxgboostspacyfaissbm25pythonjavascriptc++fastapiasynciodjangorest api designopenai apiservicenow apimicrosoft teams apichromadbazureawsdockerterraformstreamlitgitjupyterpytestblackpylintmypyisortservicenow ai predictive intelligencenow assistvirtual agentnlucmdbsap integrationentra id apimlopsdrift detection (psi)prompt engineeringclass imbalancethreshold tuningagentic ai

Work Experience

Associate System Engineer

IBM CIC

Dec 2023 - Present Pune, India
  • Predictive Intelligence: Configured category and assignment group prediction models — tuned confidence thresholds, achieving 45% reduction in ticket rerouting and 20% reduction in manual categorisation.
  • Now Assist: Configured and tested summarisation, resolution note generation, and semantic KB search — contributing to 2× faster incident resolution through LLM-powered knowledge retrieval.
  • Virtual Agent: Designed 12 NLU-driven conversational topics across ITSM and HRSD — automating self-service resolution for IT and HR service requests via catalog items and record producers.
  • API Integrations: Built real-time REST APIs with Microsoft Entra ID for dynamic group membership management and a ServiceNow KB data pipeline to feed LLM training for a Microsoft Teams AI assistant.
  • CMDB & SAP Integration: Configured CMDB, asset tables, and IDM APIs during Matrix42-to-ServiceNow migration; maintained SAP-ServiceNow HR field mappings (manager hierarchy, department) underpinning HRSD Virtual Agent personalisation.

Machine Learning Intern

IBM Labs

Jul 2023 - Sep 2023 Pune, India
  • Selected among 5 of 65 applicants for IBM Labs 2023 Innovation Programme — built ML automation tools for telecom order management.
  • Compared Random Forest vs XGBoost for failed-order auto-healing pipelines — XGBoost outperformed by 35% recall and 10% accuracy on an 8:2 imbalanced dataset using SMOTE and GridSearchCV.
  • Engineered 15–20 features from ServiceNow incident, HR case, and service request data; built a proof-of-concept auto-retry mechanism to predict order failure root causes — presented to IBM Labs stakeholders.

Backend Developer Intern

Spring Money

Jan 2023 - Jun 2023 Pune, India
  • Designed and tested RESTful APIs for a fintech mobile application, writing clean, documented Python/Django backend code aligned to customer requirements.
  • Optimised existing database schemas and queries, improving data retrieval performance for active user accounts.
  • Practised Agile/Scrum methodology across a 6-month sprint cycle in a fast-paced startup environment.

Projects

IT Incident Priority Predictor

PythonXGBoostSHAPScikit-learnFastAPIDockerStreamlit
  • Built an end-to-end ML classifier to auto-detect high-priority IT incidents — comparing Decision Tree, Random Forest, and XGBoost — achieving 0.93 recall and 0.76 F1 on a 9:1 imbalanced real-world dataset using class_weight='balanced' and threshold tuning to 0.7.
  • Engineered 18 features from raw ServiceNow incident fields including temporal patterns, escalation signals, and impact- urgency interactions; integrated SHAP explainability to surface per-prediction feature contributions.
  • Deployed FastAPI inference server with async /predict/batch endpoint using asyncio.gather() and ProcessPoolExecutor — containerised with Docker and served via Uvicorn with PYTHONPATH and environment variable injection.
  • Implemented JSONL prediction logger and PSI-based data drift detector (thresholds: <0.1 stable, 0.1–0.2 watch, >0.2 retrain) for production monitoring.
  • Deployed two-tab Streamlit app with single-ticket prediction and batch CSV upload with downloadable results.

RAG Chatbot with Hybrid Search

PythonLangChainFAISSBM25SentenceTransformersStreamlit
  • Built a full RAG pipeline from scratch — PDF ingestion → chunking → embedding (all-MiniLM-L6-v2, 384-dim) → FAISS vector store → hybrid retrieval → LLM generation — deployed as a ChatGPT-style Streamlit UI.
  • Implemented hybrid search combining FAISS semantic similarity and BM25 keyword retrieval with score-based merging for more robust retrieval than vector-only approaches.
  • Added conversational memory with query-history rewriting for context-aware multi-turn conversations across a modular architecture: ingest.py → retrieve.py → hybrid.py → generate.py → app.py.
  • Identified production gaps and improvement roadmap: RRF (Reciprocal Rank Fusion) over current order-based merge, FAISS index persistence, embedding caching, and RAGAs evaluation metrics.

AI Resume Analyzer (ATS Scorer)

PythonspaCySentence TransformersOpenAI APIStreamlit
  • Built an NLP-powered ATS resume analyser computing keyword match score, skill-weighted score, and semantic similarity between resume and job description using sentence embeddings.
  • Engineered a spaCy NLP pipeline with lemmatisation, noise filtering, and synonym-aware normalisation for robust keyword extraction beyond simple string matching.
  • Integrated OpenAI API to generate tailored resume improvement suggestions with missing keyword detection and multi- dimensional match score breakdown.

Leaderboard Standings

Leaderboard Position Pending

Global test scores, peer standing percentiles, and algorithm leaderboard ranks are updated dynamically.

Assessment Highlights

Assessments Not Completed

Coding evaluations, system assessment results, and conceptual score badges will appear here after taking a test.

AI Collaboration Score

AI Collaboration Score Pending

Developer coding behavior, assistant cooperation, and AI pair-programming indicators are evaluated during live coding sessions.

Role Compatibility Profile

Role Compatibility Analysis Pending

Custom matching reports, candidate role compatibility percentiles, and core engineer strength profiles are processed once conceptual code screenings are complete.

Achievements

Publication on AI ethics

Publication: "Overcoming Barriers and Promoting Responsible AI Development: Artificial Intelligence Ethical Considerations" — IJSRP, Vol. 13, Issue 9, Sep 2023 [ISSN 2250-3153]

About Details

Professional Bio

AI Engineer with 2.5 years of industry experience at IBM, specialising in LLM integration, RAG pipelines, and end-to-end ML systems. Proven track record in configuring predictive intelligence models and building scalable AI solutions.

B.E. Computer Science

Dr. Vishwanath Karad MIT World Peace University, Pune (2019 - 2023)

Languages: English, Hindi