// last updated Feb 2026
Senior Data Scientist
Abhijith Nair
_
Python programmer and ML practitioner with 6+ years of experience building production-grade machine learning, deep learning, and GenAI systems. Currently at Verisk.
0
years exp.
0
companies
0
AI systems
0
Stack Overflow
01
Experience
Senior Data Scientist (prior: Data Scientist I, II)
►
- Designed and deployed multiple RAG architectures using LangChain with multimodal retrievers (meetings, docs, images) for sales insights — achieving <4s streaming latency.
- Built core document AI pipelines for Underwriting Assistant, integrating vision-LLMs to segment, classify, and extract structured data from insurance documents — reducing manual workload by 30–40%.
- Deployed multimodal AI (Llama, BLIP2) for Lightspeed Small Commercial, improving image analysis precision by 15%.
- Optimized on-premise LLM inference using quantized Llama with TGI containers, reducing latency by 60% while maintaining accuracy.
- Implemented crawling pipelines for sourcing data, integrating with BIDB datasets and Senzing search UI for rapid POC iterations.
- Mentored interns and collaborated across business units on GenAI and data science initiatives.
Data Scientist → Data Scientist Intern
►
- Built an end-to-end, low-latency recommendation engine at rytfit.ai to match candidates to job opportunities.
- Implemented deep neural network models for resume segmentation and entity extraction from PDF, DOCX, and JPEG formats.
- Developed Kafka + ksqlDB + Elasticsearch data pipelines, reducing processing time by 20%.
- Deployed unsupervised GPU-clustering to infer relevant skills for AI-assisted job creation.
- Fine-tuned BERT embeddings on large resume corpora, improving candidate-job match accuracy by 14%.
- Automated data extraction and transformation from web sources using customized ETL scripts to reduce processing time.
- Deployed multiple containerized microservice APIs and quantized models into production with low latency.
Enthire, Inc
Deep Learning Intern — NLP
►
- Built a career-path prediction model using encoder-decoder sequential networks based on “Next Career Move Prediction with Contextual Embedding.”
- Engineered TensorFlow data pipelines with SQL and data wrangling for model training.
02
Skills
ML / Deep Learning
Classification & Regression
Clustering (K-Means, DBSCAN)
PCA / t-SNE / Autoencoders
Transformers / LLMs
CNNs / Segmentation
LoRA Fine-tuning
Programming
PythonSQLBash
Frameworks & Tools
PyTorchTensorFlowScikit-Learn
LangChainAutogenRagas
DockerFastAPIKafka
Cloud & DBs
AWS / GCP
PostgreSQL
ElasticsearchQdrant
Proficiency Levels
03
Projects
Multi-Adapter LLM Deployment
Fine-tuned multiple LLM models using LoRA on public prompt-completion datasets. Deployed with LoRAX for efficient multi-adapter serving — one GPU, many models.
Health QA Agent (AutoGen)
LLM-powered Q&A on health guideline documents using Microsoft AutoGen framework for complex multi-agent interactions and reasoning chains.
Synthetic Dataset Generator
Leveraged open-source instruct LLMs to generate high-quality synthetic fine-tuning data from unstructured text — improving data availability and downstream LLM performance.
04
Awards
Verisk “Way to Go” Award — exceeded expectations in Q1 & Q4 of 2023.
Verisk “Set a New Standard” — exceptional overall performance in 2022 & 2023.
IntellectFaces Best Employee of the Year — outstanding contributions in 2020.
IntellectFaces Star Performer — Q1 2021 & Q3 2020 for exceeding expectations.
Top 5% Stack Overflow — Python & Pandas tags, 2021.
Deep Learning Specialization — Coursera (2019) · Microsoft Technology Associate (2015)
Let's Build
Something Intelligent
Open to new opportunities and collaborations in AI / ML