Looking for a job

Sricharan D A

@sricharanda

Message

I build full-stack data platforms and interpretable machine learning systems.

India

Message

What I'm looking for

I'm looking to build production data products and machine learning systems where I can work across backend engineering, analytics, and interpretable AI.

I've built and deployed data-intensive platforms for Indian electoral bonds, equity earnings, and mutual-fund portfolio analytics using FastAPI, PostgreSQL, Next.js, React, and Docker.

For bonds-contracts-tracker, I connected electoral-bond disclosures to political-party redemptions with a 91.8% match rate across 18,871 purchases and 20,421 redemptions. I also built entity-resolution and scraping pipelines for donor records and 5,000 NTPC contract awards.

With earnings-edge, I processed more than one million price records and 5,800 earnings events for India's 500 largest listed companies, improving analysable coverage to 458 companies and correcting a split-adjustment issue across 836,000 rows.

As a research intern at PI Labs, PES University, I developed an interpretable multimodal mental-health screening pipeline using HuBERT and a Sparse Autoencoder. The system achieved 0.74 cross-validation AUC while identifying sparse features aligned with clinically recognised depression markers.

Experience

Work history, roles, and key accomplishments

Research Intern

PI Labs, PES University

Jun 2026 - Jul 2026 (1 month)

Designed an end-to-end multimodal pipeline integrating a pretrained HuBERT speech encoder with a custom Sparse Autoencoder, learning 6,144 sparse latent features for interpretable depression-related vocal patterns. Achieved 0.74 cross-validation AUC on the DAIC-WOZ clinical dataset, matching the raw-feature baseline while improving interpretability, validated with ablation studies and statistical

PyTorch Hugging Face Transformers Sparse Autoencoders HuBERT Machine Learning Data Analysis Python