Senior Software Engineer · Apollo Global Management

Heramb Devbhankar

I build fast, reliable backend systems and AI platforms that run in production.

5+ years shipping Python microservices, cloud-native data platforms and LLM agents, from FDA clinical trials to quant finance.

lower API latency: 60%+lower API latency
faster data pipelines: 70%faster data pipelines
lower LLM cost: 60%lower LLM cost
in production: 5+ yrsin production

Download Resume

Get in touch

Open to full-time roles · Relocating to India

What I do

Core Expertise

Backend at scale

Python microservices and APIs that stay fast under heavy load.

Cloud-native data platforms

Fault-tolerant pipelines on Kubernetes, shipped through CI/CD and GitOps.

Generative AI in production

LLM agents with MCP and text-to-SQL, tuned for latency and cost.

Experience

Where I've Made an Impact

Senior Software Engineer
Apollo Global Management · El Segundo, CA
Jan 2025 – Present
- 60%+ lower API latency
- Pipelines 13 h → 4 h
- 60% lower LLM cost
- Built FastAPI microservices on SQL Server with Redis caching and Protobuf, cutting latency 60%+ under heavy load.
- Re-architected data ingestion on Temporal: daily runs 13 h → 4 h, monthly cycles 4 days → 1.
- Rebuilt GenAI agents on MCP, tool calling and text-to-SQL, cutting LLM latency and token cost by 60%.
- Own a cloud-native data platform for real-time quant analytics (25M+ rows/year) on AKS, Terraform, Helm and GitOps.
- Python
- FastAPI
- Temporal
- Redis
- Azure AKS
- MCP
Software Development Engineer
SAG-AFTRA Foundation · Los Angeles, CA
May 2023 – Oct 2024
- Led monolith → microservices
- Led architecture and delivery of a Rails monolith to microservices migration across FastAPI, React and Swift iOS.
- Built the AWS backend and CI/CD (load balancing, async jobs), improving reliability and release frequency.
- Python
- FastAPI
- Ruby on Rails
- React
- AWS
Software Development Engineer
Eye Labs Inc · Santa Clara, CA
Sep 2022 – Apr 2023
- Baseline for FDA clinical trials
- Built gaze-tracking and facial-landmark algorithms adopted as the production baseline for FDA clinical trials.
- Shipped an iOS app (SwiftUI, Apple Vision) that detects eye anomalies from the front camera.
- Swift
- SwiftUI
- Apple Vision
- Computer Vision
Research Assistant
University of Southern California · Los Angeles, CA
Apr 2021 – Aug 2022
- Computer vision research
- Tracked movement and gene expression in fruit flies from GFP video; classified trajectories with LSTM and CNN models.
- Generated synthetic face-recognition training data with StarGAN, reducing the need for real-world data.
- Python
- TensorFlow
- OpenCV
- GANs
Software Development Engineer
IOTric · India
Aug 2019 – Dec 2020
- Built an OCR web app that extracts text from distorted invoices and receipts.
- Python
- Object Detection
- OCR

Education

M.S. Computer Science

University of Southern California

2021 – 2022

B.Tech

Dr. Babasaheb Ambedkar Technological University

2019

Projects

Featured Work

70% faster pipelines

Cloud-Native Data Platform

Powers real-time quant analytics at Apollo: 78K × 370 matrices per run, fault-tolerant Temporal pipelines, high-throughput retrieval.

Python
Temporal
SQL Server
Redis
Azure AKS

60% lower latency & token cost

GenAI Text-to-SQL Agents

LLM agents that answer data questions in plain English, built on MCP and tool calling with grounding and eval guardrails.

LLMs
MCP
Text-to-SQL
Python

Open source · PyPI

spoken2written

Open-source NLP library on PyPI that turns spoken-style text into clean written form.

pip install spoken2written

Python
spaCy
NumPy
NLP

Stanford University

Tactile Images for the Visually Impaired

Flask API, built with Stanford, that converts photos into tactile graphics using segmentation and edge detection.

Python
OpenCV
TensorFlow
Flask

Tech Stack

Tools I Ship With

Languages

Python
SQL
JavaScript
Swift
C++
C#
Ruby

Backend & Data

FastAPI
Flask
Node.js
Ruby on Rails
SQL Server
Redis
Temporal
Protobuf
Microservices

Cloud & DevOps

Azure (AKS)
AWS
Docker
Kubernetes
Terraform
Helm
GitHub Actions
Flux CD

AI / ML

LLM Agents
MCP
Text-to-SQL
TensorFlow
OpenCV
spaCy
NumPy
CNN · LSTM · GANs

Frontend & Mobile

React
Angular
SwiftUI
UIKit

Engineering

Distributed Systems
System Design
High Availability
Performance Tuning

FAQ

Quick Answers

Who is Heramb Devbhankar?

Heramb Devbhankar is a Senior Software Engineer at Apollo Global Management who builds backend systems, cloud-native data platforms and Generative AI agents. His website is heramb.tech.

What does Heramb Devbhankar specialize in?

Python backend microservices (FastAPI), distributed systems, data pipelines with Temporal, Kubernetes on Azure and AWS, and LLM agents using Model Context Protocol and text-to-SQL.

What has Heramb Devbhankar achieved?

At Apollo he cut API latency by 60%+, made data pipelines 70% faster (13 hours to 4) and lowered LLM inference cost by 60%. At Eye Labs his algorithms became the baseline for FDA clinical trials.

Where did Heramb Devbhankar study?

He earned an M.S. in Computer Science from the University of Southern California (2022) and a B.Tech from Dr. Babasaheb Ambedkar Technological University in India (2019).

How can I contact Heramb Devbhankar?

Email heramb.hvd@gmail.com, connect on LinkedIn (linkedin.com/in/heramb-devbhankar), or use the contact form at heramb.tech/contact. He is open to full-time roles and relocating to India.