PIYUSH BHAWSAR

Co-Founder

Bengaluru, Karnataka, India2 yrs 5 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Co-founded an AI startup focused on market predictions.
Designed a Generative AI curriculum for aspiring engineers.
Led a team to develop an AI customer support platform.

Stackforce AI infers this person is a Backend-focused AI Engineer with experience in customer support and analytics solutions.

Contact

Skills

Core Skills

AiGenerative AiAi AnalyticsSoftware EngineeringAi Customer SupportBackend DevelopmentAi Engineering

Other Skills

Market PredictionsFantasy SportsCurriculum DesignMulti-Agent ArchitectureSQLData ScienceFine-Tuning ModelsCachingHigh-Cardinality SamplingInstruction-RAG FrameworkKubeJob OrchestrationOpenSearchPrompt TuningDevOpsCloud Infrastructure

Experience

Sit or start ai

Co-Founder

Aug 2025 – Feb 2026 · 6 mos · San Francisco Bay Area

AI for general market, fantasy sports and trade predictions.

AIMarket PredictionsFantasy Sports

Scaler

Instructor – Generative AI

May 2025 – Jan 2026 · 8 mos · Bengaluru, Karnataka, India

Instructor & Designed the Generative AI curriculum for Scaler.

Generative AICurriculum Design

Sundial

Software Engineer

Jan 2025 – Oct 2025 · 9 mos · Bengaluru, Karnataka, India · On-site

Sundial is a three-year-old Palo Alto–based AI analytics startup providing unified analytics and AI‑driven decision intelligence, having clients like OpenAI, Character.AI, and PayPal (Venmo), with a 30‑member engineering team.
Scaled single‑agent to multi‑agent architecture (viz, SQL, data-science, classifier + NER) with dynamic routing, conditional handover, and multi‑model support (Gemini, OpenAI, Groq).
Supervised Fine‑tuned(SFT) & distilled openai models for classifier agents; built AI suggestions & typeahead, further reducing latency 2000 ms → 900 ms via caching & high‑cardinality sampling.
Designed Instruction‑RAG framework enabling detailed multi‑step analytics.
Built E2E evaluation pipeline (KubeJob orchestration, multi‑metric scoring, deterministic retrieval tests) to validate inference quality across high‑cardinality tenants.
Delivered a tenant‑configurable agent framework for model/prompt/workflow tuning, context manipulation, fallback agents, and enable/disable controls.
Enhanced retrieval accuracy with semantic + fuzzy search, reranker models, and optimized OpenSearch (n‑gram & standard analyzers).

AI AnalyticsMulti-Agent ArchitectureSQLData ScienceFine-Tuning ModelsCaching+5

Dukaan®

Backend & GenAI Lead

Jan 2024 – Dec 2024 · 11 mos · Bengaluru, Karnataka, India · On-site

Led the development of Bot9, Dukaan's AI customer support platform serving 50+ SMBs and a major enterprise, handling 2,500+ tickets and 40K+ messages daily.
Managed a team of 5 engineers, delivering 100+ features across backend and GenAI.
Cut OpenAI costs by 40% (prompt caching) and another 60% via RAG + prompt tuning.
Prototyped multi-model routing and low-latency voice support (200–300ms) on Raspberry Pi.
Scaled file uploads from 100/sec to 10K/sec using pre-signed Cloudinary URLs.
Migrated infra from WebSockets to SSE; built LLM-agnostic backend with Anthropic support.
Set up DevOps/infra: k3s, Grafana, Prometheus, OpenTelemetry, SigNoz, and serverless Lambda.

AI Customer SupportBackend DevelopmentPrompt TuningDevOpsCloud Infrastructure

Inodesoft

AI Engineer

Jul 2023 – Dec 2023 · 5 mos · New York, United States · Remote

Built a crypto market prediction pipeline using GPT-4 Vision for market graph analysis and trend estimation.
Developed a Twitter automation bot with a social network simulator by scraping real tweets, generating new ones via GPT, and simulating interactions using Langchain and Selenium.

Crypto Market PredictionGPT-4 VisionTwitter AutomationLangchainSeleniumAI Engineering