CV

Education

• Master of Science in Scientific Computing, University of Heidelberg, Germany, 2021

Work Experience

• Machine Learning Engineer, April 2022 - Present
  • GFK SE, Nuremberg
  • Deployed a monitoring and alerting system for gfkNewron with Grafana dashboard, custom metrics recording in Prometheus, and Slack integration and email notifications to achieve better microservice monitoring, data consistency, and KPI tracking.
  • Mentored junior engineers for agile process, code reviews, organizing KT sessions, and participated in interviewing.
• Data Scientist, April 2020 - May 2021
  • SAP SE, Germany (Berlin, Walldorf)
  • Fine-tuned and deployed hybrid Seq2Seq-BART model on GCP, using data augmentation, pre-training techniques for natural language-source code system.
  • Pretrained GPT-2 and RoBERTa LLM on Python code from scratch for developing RESTful web service for code auto-completion and NL2Code system.
  • Delivered an end-to-end ML PoC to business customers for automating master data categorization, sales order approval, and commodity code prediction
  • Implemented & deployed SonarQube Git Pull Request decorator in Jenkins CI/CD pipeline for static code analysis
  • Responsible for python development, improvements, bug fixing for microservice ML offering, Data Attribute Recommendation.
• Software Developer, June 2019 - March 2020
  • SAP SE, Walldorf, Germany
  • Developed clustering & embedding techniques for content classification.
  • Improved spelling correction service by integrating semantic similarity.
  • Developed lightQueue feature for crawler in NodeJS to reduce 40% RAM usage.
• Research Scholar, August 2017 - July 2018
  • Indian Institute of Technology Kanpur, India
  • Developed a neural multilingual algorithm setup to retrieve documents in language x using queries in language y in absence of aligned/parallel corpus.
  • Worked on automatic query generation for resource-starved Indian languages in absence of aligned corpus by incorporating multilingual topic modeling.
  • Supervisor: Prof. Arnab Bhattacharya

Technical Skills

• Programming: Python, Java, Bash, SQL, HTML, and LATEX

• Dev Tools: Jenkins CI/CD, Gitlab, Docker, Kubernetes, GCP, pgAdmin, Postman, REST, Scrum, Agile, Kanban

• ML Stack: PyTorch, Tensorflow, HuggingFace, Sklearn, Pandas, NumPy, FastAI, spaCy, NLTK, Gensim, VertexAI, Jupyter, LangChain, MLflow Airflow, Streamlit, SHAP, ONNX, DeepSpeed

Language Skills

• Languages: English (Fully Proficient), German (B1.1), Bengali(Native), and Hindi (Advanced)

Strengths

1. Proficient in breaking down business problems into machine learning solutions using ML stack (Pytorch, Scikit-learn, Tensorflow, HuggingFace, MLFlow).

2. Python backend development (deploying new features on SaaS, migrating codebase from Tensorflow v1.0 to v2.0, etc.) and DevOps (Kubernetes, GCP, Docker, Jenkins CI/CD).
3. Cross collaboration in agile teams with stakeholders and acted as a bridge and managed a smooth workflow between research and engineering teams, and ability to organise KT sessions on machine learning and software engineering principles across the company.

Publications

IR-IITBHU at TREC 2016 Open Search Track: Retrieving documents using Divergence From Randomness model in Terrier

M. Niyogi, S. Pal. (2017). "IR-IITBHU at TREC 2016 Open Search Track: Retrieving documents using Divergence From Randomness model in Terrier." The Twenty-Fifth Text REtrieval Conference (TREC 2016) Proceedings. NIST Special Publication:SP 500-321.

Business: Do you wanna sell more? Discovering Topics, Sentiments and Prediction of Ratings

Niyogi, M., Kumar Pal, A. (2019). "Business: Do you wanna sell more? Discovering Topics, Sentiments and Prediction of Ratings." In Proceedings of the 7th International Conference on Computer and Communication Technology (ICCCT-2017). Association for Computing Machinery, New York, NY, USA, 133–138.

Learning Multilingual Embeddings for Cross-Lingual Information Retrieval in the Presence of Topically Aligned Corpora

M. Niyogi, K. Ghosh, and A. Bhattacharya. (2018). "Learning Multilingual Embeddings for Cross-Lingual Information Retrieval in the Presence of Topically Aligned Corpora." arXiv preprint arXiv:1804.04475v1.

Discovering Conversational Topics and Emotions Associated with Demonetization Tweets in India

Niyogi, M., Kumar Pal, A. (2019). "Discovering Conversational Topics and Emotions Associated with Demonetization Tweets in India." In: Verma, N., Ghosh, A. (eds) Computational Intelligence: Theories, Applications and Future Directions - Volume I. Advances in Intelligent Systems and Computing, vol 798. Springer, Singapore.

Neural Models for Source Code Synthesis and Completion

Mitodru Niyogi. (2021). "Neural Models for Source Code Synthesis and Completion." IWR, Heidelberg University, Germany.

Talks

Artist identification and generating new paintings using DC-GAN

August 05, 2019

Talk at Faculty of Mathematics and Computer Science, Heidelberg University, Heidelberg, Germany

Special Topics in Continuous Optimization and Optimal Control: Local and Global convergence of Newton methods

December 07, 2019

Talk at Faculty of Mathematics and Computer Science, Heidelberg University, Heidelberg, Germany

ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Language Tasks

January 15, 2020

Tutorial at Faculty of Computational Linguistics, Heidelberg University, Heidelberg, Germany

Review of Faithful to the Original: Fact Aware Neural Abstractive Summarization: Ziqiang et al.

January 22, 2020

Talk at Faculty of Computational Linguistics, Heidelberg University, Heidelberg, Germany

Review of Deep Recurrent Neural Networks for Prostate Cancer Detection: Analysis of Temporal Enhanced Ultrasound (Shekoofeh Aziz et al.)

June 25, 2020

Talk at Interdisciplinary Center of Scientific Computing, Heidelberg University, Heidelberg, Germany

Neural Code Synthesis and Completion

August 25, 2021

Talk at Faculty of Mathematics and Computer Science, Heidelberg University, Heidelberg, Germany

Paper Review: Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank

October 03, 2022

Talk at TU Darmstadt, Darmstadt, Germany

Research presentation at UKP Lab, TU Darmstadt

November 21, 2022

Talk at UKP, TU Darmstadt, Darmstadt, Germany

Tutorial on Feature Importance

June 23, 2023

Talk at Turo, Paris, Paris, France

Tutorial

Service and leadership