Harsh Soni

Co-Founder

3 yrs 9 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

Architected real-time data pipelines for 10M+ events/day.
Contributed to major open-source projects like TensorFlow.
Achieved significant performance improvements in backend systems.

Stackforce AI infers this person is a Backend-focused Software Engineer with expertise in Distributed Systems and Data Engineering.

Contact

Skills

Core Skills

Distributed SystemsBackend DevelopmentData EngineeringWeb DevelopmentSoftware EngineeringOpen Source Contribution

Other Skills

Spring BootApache KafkaScyllaDBKubernetesPythonDjangoPySparkAirflowKafka StreamsRedisDockerGitHubPHPJavaScriptWordPress

About

I’m a Software Engineer who loves designing scalable backend systems, building reliable infrastructure, and solving hard engineering problems. At my current role as a founding engineer, I’ve architected real-time pipelines that handle 10M+ events/day across 10k+ edge devices, built distributed services in Python and Java, and shipped data infrastructure used for analytics, telemetry, and ML feature generation. I enjoy working across the stack, from designing clean APIs and distributed storage paths, to optimizing data workflows, to automating deployments with Kubernetes and GitOps. I’m especially passionate about system performance, correctness, and elegant abstraction design. I actively contribute to open-source, including TensorFlow Datasets (Google) and scikit-learn, where I’ve worked on dataset tooling, video pipelines, and preprocessing optimizations used by thousands of developers. With 1300+ DSA problems solved, a 9.63 GPA, and experience owning end-to-end systems from 0→1, I’m driven by engineering excellence and building software that scales in the real world. If you're working on challenging distributed systems, infra platforms, ML/data pipelines, or anything that pushes technical boundaries — I’d love to connect.

Experience

3 yrs 9 mos

Total Experience

1 yr 10 mos

Average Tenure

3 yrs 4 mos

Current Experience

Six sense mobility

Founding Engineer

Jan 2023 – Present · 3 yrs 4 mos · Delhi, India · Hybrid

Led the development of a connected car platform, enabling real-time monitoring of vehicle parameters.
Architected an event-driven data ingestion pipeline using Kafka Streams and ScyllaDB, achieving 99.9% uptime while processing 10M+ records/day from 10k+ edge devices.
Designed and deployed modular Spring Boot and Django microservices, reducing feature roll-out time by 35% using SOLID clean architecture principles.
Engineered a low-latency telemetry backend (sub-10ms write latency) with ScyllaDB Redis, enabling real-time insights for fleet monitoring and predictive maintenance.
Developed PySpark + Airflow pipelines to batch-process terabytes of telemetry data, accelerating analytics by 3x and powering downstream ML models.
Built OTA firmware delivery with JFrog Artifactory, reducing distribution time by 40% across 1000s of vehicles.
Set up observability stack (Prometheus + Grafana + AlertManager), cutting incident resolution time by 50%.
Pioneered infrastructure-as-code with ArgoCD and Kubernetes, enabling GitOps deployment across staging and prod.

Spring BootApache KafkaScyllaDBKubernetesPythonDjango+4

Shipsy

Software Engineering Intern

Jan 2022 – Jun 2022 · 5 mos · Gurugram, Haryana, India

Spearheaded the development of a WordPress plugin to integrate WordPress shops with the organization’s system.
Developed efficient cron jobs to reduce clicks required for order syncing by an average of ∼54% for customers.
Reduced debugging time by ∼11% by building a request logging system with the decorator pattern.
Improved application speed by implementing singleton pattern to prevent multiple instances of DB interactions.
Streamlined development process and reduced API calls by ∼22% by building internal APIs for seamless data and function sharing between PHP and JavaScript.
Dockerized the application and WordPress to enable fast development and hassle-free setup.
Cut upfront setup time by ∼67% by creating a custom Docker image for Magento, simplifying local setup.
Automated the upload of a configured plugin to Slack using a custom-built Slack command and Jenkins pipeline, reducing dependency and enabling quick application delivery.
Reduced application approval time by ∼54% by developing GitHub actions for code sniffing based on WPCS.

DockerGitHubPHPJavaScriptWeb DevelopmentSoftware Engineering

Tensorflow

Open Source Contributor (TFDS)

Oct 2020 – Jan 2021 · 3 mos

Contributed to TensorFlow Datasets library.
Remodeled TFDS's dataframe visualization for bounding boxes and videos. Restrict the display of Sequence/batch elements to single line by enabling scrolling.
Wrote a script to record filenames in checksums for both legacy and new datasets.
Developed a feature to allow users to skip downloading datasets and use manually downloaded datasets.

GitHubPythonOpen Source Contribution

Scikit-learn

Open Source Contributor

Jul 2019 – Sep 2020 · 1 yr 2 mos

Contributed to scikit-learn
Optimized MissingIndicator implementation by computing mask only once which reduced runtime by ~0.5%
Enhanced code to use the precomputed mask in MissingIndicator for both sparse and dense matrices.

GitHubPythonOpen Source Contribution