---
title: "Abhishek Saxena - Member of Technical Staff - 1 in Bengaluru | Tal"
description: "Meet Abhishek Saxena, Backend Engineer @ Retape.ai (YC W23) | Go · Kafka · Distributed Systems | Ex-SuperAGI, Probo | NIT KKR ’25 based in Bengaluru. AI stack includes claude code, codex, and mcp. Explore public work links including github and leetcode."
image: "https://tal.club/meet/abhishek-saxena-4/share-image"
---

# Abhishek Saxena on Tal

## Headline

Backend Engineer @ Retape\.ai \(YC W23\) \| Go · Kafka · Distributed Systems \| Ex\-SuperAGI, Probo \| NIT KKR ’25

## About

Backend engineer who builds data\-intensive systems that stay fast and correct under load — high\-throughput pipelines, zero\-downtime migrations, and real\-time sync across MySQL, MongoDB, OpenSearch, Clickhouse, and Neo4j\. Currently MTS at Retape AI \(YC W23\), working at the intersection of applied LLMs and infrastructure\. Previously shipped scalable Go services at SuperAGI and cut API latency 85% at Probo\.

## Location

Bengaluru

## AI stack

claude code

codex

mcp

ollama

Zapier

## Work experience

Member of Technical Staff \- 1 at Retape AI, Bengaluru, Customer Analytics: Built a production analytics platform processing 3\.7M\+ records across 66 creditors, with automated recomputation and 14 fail\-closed invariants achieving zero discrepancies\. Scrub Analyzer: Built an autonomous data\-quality watchdog combining deterministic anomaly detection with an LLM SQL agent, reducing false\-positive alerts by 78% across 10\+ production environments\. Creditor Resolution: Built a deterministic \+ AI hybrid to resolve free\-text creditor notices, handling \~90% of traffic instantly and achieving zero bad writes since launch\.

Software Engineer at SuperAGI, Sep 2025 \- May 2026 · 9 mos, Gurugram, Legacy Monolith to Microservice Migration • Executed zero\-downtime migration of 5M\+ records from a legacy Ruby monolith to a Go\-based views microservice serving 1000\+ requests/secs\. • Implemented dual\-write based real\-time sync using model concerns and migration APIs to maintain consistency during rollout\. • Migrated large\-scale historical data using Sidekiq batches and background jobs, ensuring 0 data loss and 0 production issues\. Generic Webhook Delivery Infrastructure • Built a high\-throughput generic webhook service in Go handling 1000\+ events/min across multiple integrations\. • Developed webhook ingestion endpoints to trigger internal automation workflows and event\-driven processes\. • Implemented reliable delivery using exponential backoff retry strategies for failed webhook attempts\. • Added SSRF protection and request validation to secure outbound webhook execution\. Bulk Sync & Distributed Data Consistency • Built large\-scale bulk update/delete workflows for CRM objects during migration from MySQL\-based legacy services to MongoDB\-backed CRM services\. • Designed Sidekiq\-based distributed batch pipelines that split operations into batches of 1000 records for scalable execution\. • Triggered Kafka events from each batch to synchronize MongoDB services with legacy MySQL tables in near real time\. • Published MQTT completion events after successful batch execution to drive downstream workflow orchestration and tracking\.

Software Engineer at Probo, Jun 2025 \- Sep 2025 · 3 mo, Gurugram, Entity Resolution & Semantic Retrieval Systems • Built a Python\-based entity resolution system to identify the correct entity from ambiguous queries \(e\.g\. multiple players with the same name\) using graph relationships and metadata signals\. • Designed Neo4j\-based relation scoring using shared matches, common news associations, connection counts, and entity attributes such as player roles, achieving sub\-100ms query latency\. • Implemented configurable frequency and recency\-based ranking signals to improve relevance for time\-sensitive queries\. • Optimized retrieval architecture by shifting from event\-centric to entity\-centric indexing, significantly improving search performance and scalability\. Low\-Latency Entity Search Optimization • Replaced FAISS\-based vector search used for entity identification with OpenSearch fuzzy search to improve query performance for ambiguous user inputs\. • Reduced entity lookup latency from 2–3 seconds to sub\-100ms while maintaining retrieval accuracy through graph\-based entity resolution signals\. • Built configurable semantic search pipelines with feature\-flag based rollout support for OpenSearch vector search experimentation and gradual migration\. • Improved scalability and search responsiveness by shifting from local vector search pipelines to optimized OpenSearch indexing and retrieval strategies\. Real\-Time Sports Data Processing Pipelines • Built RabbitMQ\-based bulk ingestion pipelines to process real\-time cricket vendor events including ball\-by\-ball updates, wickets, and match summaries\. • Designed idempotent OpenSearch update strategies using deterministic \`key\_id\` generation for consistent event processing\. • Ensured data consistency during out\-of\-order event delivery by updating records only when incoming event timestamps were newer than existing data\.

Software Intern at Probo, Jan 2024 \- May 2025 · 1 y 4 mo, Gurugram, Haryana, India, • Built a generic fetcher API that generated dynamic Elasticsearch queries from user filters, reducing query implementation time by 70% and unifying access across data domains\. • Designed a data consumer using bulk writes to reduce DB calls by 90%, eliminated read operations with stateless transformers, and ensured data consistency using timestamp validation\. • Implemented backend support for views, likes, and multi\-level comments, including edge case handling for anonymous users\. • Developed a robust socket client for real\-time data retrieval using a pub/sub architecture and Redis, reducing API response time from 350 ms to 50 ms, resulting in an 85\.7% improvement\. • Integrated vernacular support into the application feed, enabling users to interact with content in 5\+ languages and increasing user engagement by 20%\. • Built a WebSocket server capable of handling up to 3000 messages per second, supporting large\-scale, real\-time data streaming applications\. • Designed and developed an intuitive front\-end and back\-end admin dashboard, optimizing user and contract management processes, and reducing task completion time by 30%\.

Web Development Intern at BharatOne Care, Oct 2023 \- Dec 2023 · 3 mos, Remote, \- Designed a reusable backend\-driven table component used across the codebase, streamlining table generation and reducing development time\. \- Created a JSON\-driven form generator component using React\.js, standardizing form creation across the entire codebase and accelerating development\. \- Implemented several API using Express\.js and utilized Firebase Admin SDK for data storage in Firestore, enhancing data management\. \- developed the facilitator dashboard allowing facilitators to easily access agents information\. \- developed on\-boarding for agents, hospitals and facilitators

## Education

Bachelor of Technology \- BTech, Electronics and Communications Engineering, National Institute of Technology, Kurukshetra, Haryana

## Public links

- github: https://github\.com/abhishekjedi
- leetcode: https://leetcode\.com/opt1mus\_01

[View the original profile](https://tal.club/meet/abhishek-saxena-4)

```json
{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "Person",
      "@id": "https://tal.club/meet/abhishek-saxena-4#person",
      "name": "Abhishek Saxena",
      "url": "https://tal.club/meet/abhishek-saxena-4",
      "image": "https://tal.club/meet/abhishek-saxena-4/profile-photo.jpg",
      "description": "Meet Abhishek Saxena, Backend Engineer @ Retape.ai (YC W23) | Go · Kafka · Distributed Systems | Ex-SuperAGI, Probo | NIT KKR ’25 based in Bengaluru. AI stack includes claude code, codex, and mcp. Explore public work links including github and leetcode.",
      "jobTitle": "Member of Technical Staff - 1",
      "worksFor": {
        "@type": "Organization",
        "name": "Retape AI"
      },
      "hasOccupation": {
        "@type": "Occupation",
        "name": "Member of Technical Staff - 1"
      },
      "homeLocation": {
        "@type": "Place",
        "name": "Bengaluru"
      },
      "alumniOf": [
        {
          "@type": "EducationalOrganization",
          "name": "National Institute of Technology, Kurukshetra, Haryana"
        }
      ],
      "knowsAbout": [
        "claude code",
        "codex",
        "mcp",
        "ollama",
        "Zapier"
      ],
      "sameAs": [
        "https://github.com/abhishekjedi",
        "https://leetcode.com/opt1mus_01"
      ],
      "mainEntityOfPage": {
        "@id": "https://tal.club/meet/abhishek-saxena-4#webpage"
      }
    },
    {
      "@type": "ProfilePage",
      "@id": "https://tal.club/meet/abhishek-saxena-4#webpage",
      "url": "https://tal.club/meet/abhishek-saxena-4",
      "name": "Abhishek Saxena on Tal",
      "description": "Meet Abhishek Saxena, Backend Engineer @ Retape.ai (YC W23) | Go · Kafka · Distributed Systems | Ex-SuperAGI, Probo | NIT KKR ’25 based in Bengaluru. AI stack includes claude code, codex, and mcp. Explore public work links including github and leetcode.",
      "dateModified": "2026-09-21T05:22:06.055Z",
      "isPartOf": {
        "@id": "https://tal.club/#website"
      },
      "mainEntity": {
        "@id": "https://tal.club/meet/abhishek-saxena-4#person"
      },
      "primaryImageOfPage": {
        "@type": "ImageObject",
        "url": "https://tal.club/meet/abhishek-saxena-4/profile-photo.jpg",
        "contentUrl": "https://tal.club/meet/abhishek-saxena-4/profile-photo.jpg",
        "caption": "Abhishek Saxena profile photo on Tal"
      },
      "inLanguage": "en-IN"
    }
  ]
}
```