gsasikiran / transformers

🤗Transformers: State-of-the-art Natural Language Processing for Pytorch and TensorFlow 2.0.

Hello 👋,

I am Sasi Kiran. I am data scientist interested in Natural Language Applications!

You can reach me out at my mail or LinkedIn.

Experience 💼

• Data Scientist at Ensun GmbH, Germany in Semantic search and Keyphrase Extraction

  • Semantic search algorithm for tech-scouting
  • Generating semantic keyphrases for a given text
  • Development and deployment of algorithm

• Research assistant at Fraunhofer FKIE, Germany in sub-ontology generation from natural language query.

  • Designing and implementation of semantic based sub-ontology generation for a natural language search query
  • Generation of custom Word2Vec, GloVe and FastText word embeddings for the terms in the ontology
  • Implementation and evaluation of pretrained transfer learning models BERT, ELMo
  • Collecting and processing the domain-specific text data from Wikipedia for generalization of search queries

---

Thesis 📘

• Automatic formative assessment of students’ short text answers through feature extraction
  • Generate unsupervised automated feedback for students' short text answers based on the evaluator's question and desired answer
  • Provide feedback on interchange of topics, missed topics, missed terms, and irrelevant terms in students' answers
  • Use of natural language processing features,
    • Topic modeling
    • Information retrieval
    • Coreference resolution
    • Keyphrase extraction
    • Semantic similarity
  • Automatic and manual evaluation of the formative feedback system

---

Other projects 👨🏻‍💻

• Automatic question generation
• Semantic analysis of BERT, ELMo, GPT
• Sensor fusion application for extraction and combining results from different sensors
• Classification of machine tools dataset through Transfer learning
• Monitoring neuron activations
• Live skin detector
  • Real time skin detector trained on PCA, MLP and SVM
  • Data taken from spectroscopy results

---

Research 🔬 and Teaching 👨‍🏫

• Comparative Evaluation of Pretrained Transfer Learning Models on Automatic Short Answer Grading(ArXiv pre-print)
• A brief summaries of top 30 articles of deep learning (as of May 2019)
• I also love teaching and guiding besides learning new technologies
  • Stanford CIP Section Leader: Teaching mentor for programming methodologies in Python (April 2021)

---

Formal education 👨🏻‍🎓

• Masters in Autonomous Systems from Bonn-Rhein-Sieg University of Applied Science, Germany
  • Course Work: Natural language processing, Machine learning, Neural networks, Artificial intelligence, Advanced software technology, Scientific working, Mathematics for robotics and control
• Undergraduation in Mechanical Engineering from V.R.Siddhartha Engineering College, India

---