Looking for a job

Shashank Navad

@shashanknavad

Software Engineer building production-grade pipelines, real-time data platforms, and low-latency AI systems for scalable applications.

India

Message

What I'm looking for

I’m looking for a team where I can own end-to-end data & ML systems—building scalable pipelines, LLM/RAG and real-time lakehouses—turning models into reliable low-latency production services with strong engineering rigor.

I’m a Software Engineer focused on real-time distributed systems and production pipelines. I bring expertise in Spark, Delta Lake, Kafka, and Kubernetes, and have built low-latency retrieval and ingestion systems achieving sub-200ms performance for real-time applications.

Most recently, I built a BERT-based job-role matching model that uses custom quantile loss and oversampling to produce calibrated 0–100 scores, eliminating scoring bias. I also designed a hybrid inference pipeline with in-house BERT embeddings plus a fine-tuned OpenAI re-ranker, increasing placement precision by 12% while reducing mismatches, and I productionized the system by containerizing FastAPI services on AWS Fargate.

Earlier, at PricewaterhouseCoopers, I deployed a scalable resume screening solution on Microsoft Azure using AKS—automating scalable Go microservices with Docker to handle 500+ concurrent requests with sub-3-second response times. I’m also driven by real-time data engineering, having engineered Medallion Lakehouse pipelines in Spark/Delta with exactly-once semantics and time-travel recovery, plus agentic RAG and LLMOps workflows (LangGraph/LangChain) that cut inference latency by 35%.

Experience

Work history, roles, and key accomplishments

Software Engineer - Data ML

W4M.AI

Jul 2025 - Oct 2025 (3 months)

Built a BERT-based job-role matching model with calibrated 0–100 scoring by applying custom quantile loss and oversampling to reduce scoring bias. Designed a hybrid BERT embedding + OpenAI re-ranker inference pipeline that improved placement precision by 12% and reduced mismatches, then containerized FastAPI on AWS Fargate and optimized React/TypeScript for 25% faster UI performance.

BERT OpenAI Reranker fastAPI AWS Fargate React TypeScript Docker Quantile Loss Oversampling

Technology Consultant

PricewaterhouseCoopers

Dec 2022 - Jul 2023 (7 months)

Deployed a scalable resume screening solution on Microsoft Azure using AKS, running containerized Go microservices that handled 500+ concurrent requests with sub-3-second response times. Improved recruitment efficiency by 60% by engineering an NLP-powered screening microservice with TF-IDF and lemmatization, reducing manual query screening time by 8 hours per day.

Azure Container Service Docker Microservices Azure NLP TF IDF Lemmatization Performance Tuning

Machine Learning Intern

Corizo

May 2022 - Jul 2022 (2 months)

Developed a music genre and sentiment classification prototype using feedforward networks and BERT sentiment analysis, achieving a 25% engagement lift in feedback testing. Automated lyrics ingestion via a Selenium web-scraping pipeline to collect 10 years of Billboard Hot 100 data (1,000+ songs) and eliminate manual collection.

BERT Sentiment Analysis Music Genre Classification Machine Learning Web Scraping Selenium Data Ingestion Model Evaluation