AI/GenAI Engineer · Award-Winning Innovator · Research & Production Impact · LLMs • RAG • Agentic AI • OSS Contributor
"I build robust, scalable AI/GenAI and agentic solutions—bridging R&D and enterprise, and making knowledge accessible to all."
- 🥇 1st Place, AI Hiring Show by Rabbitt AI: Led winning team (500+) for "BachatBot", an LLM financial assistant.
- 🥉 2nd Runner-Up, Hire-A-Thon (Geek Room & InvoLead)
- 🥉 SkyHack 2.0 Winner (Top 3 of 1600+ among DTU,IGDTUW,NSUT,PEC,PLAKSHA,DU)
- 🏅 5th Place, Data Engineering Summit 2025: E-commerce Product Rating ML/DataOps Challenge.
- 🏆 Top 10 (international) in:
- LT-EDI@LDK 2025 (Misogyny Meme Detection, multimodal/shared task)
- FIRE 2025 DravidianCodeMix (Offensive Language in Indian code-mixed data)
- 💡 Kaggle: Top 100/2,000 (NIFTY50 Volatility); 17th/600+ (SHL Grammar Scoring, BERT on speech)
- 🥇 Keploy API Fellow ‘25: (Top 1,000/18,500+)
- 🛡️ UGC NET (Top 6%), AFCAT + SSB Qualified
- 💻 400+ Leetcode/code problems solved
- Data Science Intern (GenAI), InvoLead:
- Led GenAI research (Knowledge Distillation, RAG, Embeddings); published internal whitepaper; built/deployed scalable ML pipelines for pharma/enterprise clients using AWS, MongoDB, Docker.
- Machine Learning Intern, IBM (CSRBOX):
- Built SVR models for academic prediction; applied analytics for SDG-focused projects.
- Data Analyst, Dusker AI:
- Automated data ETL and reporting; scaled SQL/MySQL analytics on live education data, delivering insights to product leads.
- Data Science Intern, CodSoft:
- Delivered data cleaning, preprocessing, and modeling for real-world DS projects.
- RAGBot (DUCS): Academic chatbot with FAISS, SBERT, and Gemini—boosted Q&A accuracy for 500+ users.
- MCP Resume Server: Agentic pipeline: parses GitHub, LLM-powered summaries, auto-updates JSON Resume.
- Synapse AI Hackathon Suite: Awarded multi-agent system for profile parsing, extraction, and intelligent automation.
- Voices-Reimagined: Built prize-winning real-time S2S pipeline (speech–>text–>summary–>speech, 900+ live demo).
- Suicidal Intention Detection: Created a BERT-powered pipeline for classifying suicide risk (90%+ on 232k posts), enabled mental health research impact.
Find full portfolio and presentations: datascienceportfol.io/beloabhigyan
- Sentence Transformers, UKPLab: Mainlined critical fix for JSON serialization (Python 3.12+ bugs), enhancing library for global devs.
- Shared Tasks (peer-reviewed):
- LT-EDI@LDK 2025: Top 10 global, Misogyny Meme Detection (deep learning, multimodality)
- FIRE 2025 DravidianCodeMix: Top 10, offensive lang. detection for code-mixed Indian languages (team DUCS)
- CLMIR 2025: Crosslingual Math IR, advancing info access for low-resource Indic languages
- Low-Resource Indic Translation: NMT systems for resource-constrained settings.
- System reports, posters, and collaborative workshops presented at FIRE/LDK/CLMIR; contributed to peer-discussion and methods docs.
Programming | ML / NLP & GenAI | Data / DB | Cloud / Deployment | Developer / Agentic Tools |
---|---|---|---|---|
Python, SQL, C, C++, Java, JavaScript, Node.js | HuggingFace Transformers, PyTorch, TensorFlow, Scikit-learn, BERT, FAISS, LangChain, OpenCV, SpeechBrain | MongoDB, MySQL, Pandas, NumPy, Excel, Seaborn, Plotly | AWS (S3, EC2, Batch, ECR), Docker, Streamlit, Containerization | Git(Hub), VS Code, PyCharm, Jupyter, CrewAI, Prompt Engineering, API Integration, MCP |
- Machine Learning with Python (freeCodeCamp)
- IAB Digital Marketing & Media (Google)
- IoT (Stanford)
- Arduino ATMega (MoE-IIC/DU)
- Cybersecurity (Cisco)
Open to:
- AI/ML, GenAI, NLP engineering, research, fellowships, and open-source projects.
- Collaborations at the intersection of LLMs, RAG, Agentic AI, and social impact.
Contact:
"Building AI that is robust, explainable, and transformative—bringing advanced technology to every language and user worldwide."