Software Engineer

Hi, I'm Pranay Das.

I build the distributed systems and high-throughput APIs that fintech and pharma companies trust with regulated data — automating regulatory pipelines at Eli Lilly, scaling Instoried to 100K+ users as a founding engineer, and shipping freelance KYC/AML and compliance platforms for Natixis and Novo Nordisk.

📍 San Francisco, CA 7+ years experience
Get in touch Download Resume GitHub
Distributed Systems FinTech Open Source

TOP 3% TALENT

Vetted byHire me

Google Summer of Code 2024 alum

About

Building systems that hold up under load

I'm a software engineer with 7+ years of experience building scalable distributed systems, microservices, and high-performance APIs across FinTech and pharma — at Eli Lilly, as a founding engineer at Instoried, and as a freelancer for Natixis and Novo Nordisk during my master's. I work primarily in Python, Java, and C/C++, with hands-on experience in FastAPI, Django, Temporal, Kafka, and cloud-native architectures on AWS and GCP using Docker and Kubernetes. I'm an open-source contributor with a strong systems and security mindset, and I hold a Bachelor's in Computer Science from Sambalpur University and an MS in Computer Science (Cyber Security) from Purdue University.

Software Engineering Distributed Systems Cloud Infrastructure Event-Driven Systems Platform Engineering Applied AI

Distributed Systems

Microservices, Kafka, Temporal, event-driven architecture, concurrency & caching at scale

Backend & APIs

Python, FastAPI, Flask, Django, REST, gRPC, GraphQL, OpenAPI

Cloud & Infra

AWS, GCP, Docker, Kubernetes, Terraform, Jenkins, CI/CD

Experience

Where I've worked

Software Engineer @ Eli Lilly and Company

Aug 2025 — Present · Remote
  • Proactively wrote the design doc and built a slot reservation service for radiopharmaceutical dose delivery that absorbs a 3K+ requests/sec morning rush as 10K+ clinical technicians race for early slots; idempotency keys drop duplicate retries and optimistic concurrency control eliminates double bookings.
  • Built a patient context service for adverse event classification that retrieves live medical history and redacts PII before model input, serving context in ~3 ms across 8K+ cases/day.
  • Built a citation verification layer for LLM-based adverse event classification that deterministically checks each citation exists and is permitted in model context, then uses a small model to confirm it supports the claim, letting ~70% of cases (~5.6K/day) bypass manual review.
  • Built Python transformers mapping 8K+ cases/day to MuleSoft and Veeva formats; reviewed ~20 PRs for downstream stages.

Software Engineer @ Google Summer of Code

May 2024 — Aug 2024 · Remote
  • Built and published scorecode, a Python package that retrieves OpenSSF Scorecard supply-chain security metrics; adopted as a ScanCode.io dependency with ~2K monthly PyPI downloads.
  • Integrated Scorecard results into ScanCode.io through reusable Django models, adding supply-chain security scores to its CycloneDX SBOM exports, with pytest coverage for public and private repositories.
  • Iterated with maintainers through code reviews, refining API design, naming, documentation, and code quality.

Software Engineer @ Instoried Founding Engineer

Aug 2019 — May 2023 · Bengaluru, India
  • Led a team of 5–6 engineers delivering DevOps automation, monitoring, and authentication/authorization features; hired ~5 engineers and drove technical design, code reviews, and mentoring alongside the CTO.
  • Built the backend from scratch (Python, FastAPI, PostgreSQL) for an AI writing assistant recommending tone, emotion, and concision improvements, scaling to 100K+ registered users and 30K+ daily actives.
  • Built an OAuth 2.0 service and developer portal letting ~20 B2B customers embed our recommendation APIs, driving ~20% of revenue.
  • Designed a Git-like text versioning system using snapshots and incremental deltas that sends only changed chunks on every save, cutting payloads from up to ~10 MB to a few KB and letting any lost chunk be restored from the last snapshot.
  • Resolved serving failures under 3× campaign traffic spikes by adding Redis caching (content-hashed keys, TTL, LFU eviction) and Kubernetes horizontal pod autoscaling, cutting repeat recommendation latency from 1–2 s to ~1 ms.
  • Built a real-time sales dashboard from scratch, fed by an incremental, checkpoint-based ETL pipeline that synced onboarding and campaign data without adding load to the primary database.

Software Developer @ KaptureCRM

Jan 2019 — May 2019 · Bengaluru, India
  • Designed a bidirectional sync backend connecting Shopify, HubSpot, NetSuite, Snowflake, and SQL databases to the CRM through a canonical object model, so each source needed only its own mapper, supporting 50+ onboarded clients.
  • Matched change capture to each source type (webhooks for SaaS platforms, WAL-based CDC for SQL databases, and offset-based polling workers for legacy systems), propagating changes in both directions in near real time with eventual consistency.

Software Developer @ Aptus Data Labs

Jun 2017 — Dec 2018 · Bengaluru, India
  • Built 5 REST API connectors, mostly sources such as QuickBooks and Zendesk, for an early-stage data migration platform that moved user-authorized data, limited to granted scopes, into destinations such as PostgreSQL.
  • Built Java-based Azure and JSON data-source plugins, letting Yellowfin BI users visualize that data directly.

Projects

Selected work

Open Source · GSoC '24

scorecode

PyPI package (~2K monthly downloads, a ScanCode.io dependency) that retrieves software supply-chain security metrics from the OpenSSF Scorecard API, with Django models and a ScanCode.io integration that include Scorecard results in CycloneDX SBOM exports.

GitHub PyPI
Freelance · Natixis Investments

KYC Portal

KYC/AML screening platform (Python, FastAPI, PostgreSQL) that replaced spreadsheet workflows and cut screening and review turnaround from about a week to 8 hours. A Kafka pipeline pulls data from Moody's, Refinitiv World-Check, and internal banking APIs in parallel, with real-time progress over Server-Sent Events and SSO-based access.

Proprietary client project — no public repository

Freelance · Novo Nordisk

Compliance Checker

Python/FastAPI backend that extracts information from clinical PDFs, checks it against client-defined compliance rules, flags violations, and applies corrections. Replaced two always-on 32 GB EC2 instances with on-demand parallel AWS Lambda functions that split 400-page documents into 20–30 page tasks (6–10 s each), cutting compute cost by ~80%, with DynamoDB aggregation, automatic retries, and dead-letter queues.

Proprietary client project — no public repository

Skills

Tools & technologies

Languages

Python, Java, C/C++, SQL, Bash

AI/ML

LLMs, RAG, Prompt Engineering, LLM Evaluation, Claude Code, Copilot

Frameworks & APIs

FastAPI, Flask, Django, REST APIs, gRPC, GraphQL, OpenAPI, pytest

Cloud & DevOps

AWS (S3, Lambda), Google Cloud Platform, Docker, Kubernetes, Terraform, Jenkins, GitHub Actions, CI/CD

Databases

PostgreSQL, MySQL, MongoDB, Redis, Cassandra, Elasticsearch, DynamoDB

Distributed Systems & Observability

Apache Kafka, Temporal, Microservices, Prometheus, Grafana

Education

MS Computer Science (Cyber Security) — Purdue University (2023–2025)
Bachelor's, Computer Science — Sambalpur University (2013–2017)

Contact

Let's build something

Open to Software Engineer, Backend Engineer, Platform Engineer, and Distributed Systems Engineer roles — DM me on LinkedIn or email below.

Email GitHub LinkedIn Toptal